Merge pull request #17 from Bizonu/master

Some coding style improvement, warnings disabling option

1 files changed, 3860 insertions, 3858 deletions

diff --git a/NEON_2_SSE.h b/NEON_2_SSE.h
index 7908508..c7a8c34 100644
--- a/NEON_2_SSE.h
+++ b/NEON_2_SSE.h
@@ -43,13 +43,13 @@
 #define NEON2SSE_H
 
 /*********************************************************************************************************************/
-//!!!!!!!!!!!!!! 
+//!!!!!!!!!!!!!!
 //if USE_SSE4 is defined, some functions use SSE4 instructions instead of earlier SSE versions, when undefined - SIMD up to SSSE3 are used
 //For older devices without SSE4 support it should be undefined,  for newer devices - defined, probably manualy if your compiler doesn't set __SSE4_2__ predefine
 #ifndef USE_SSE4
-#if defined(__SSE4_2__)
-    #define USE_SSE4
-#endif
+#   if defined(__SSE4_2__)
+#       define USE_SSE4
+#   endif
 #endif
 /*********************************************************************************************************************/
 
@@ -58,8 +58,8 @@
 #include <pmmintrin.h>     //SSE3
 #include <tmmintrin.h>     //SSSE3
 #ifdef USE_SSE4
-#include <smmintrin.h> //SSE4.1
-#include <nmmintrin.h> //SSE4.2
+#   include <smmintrin.h> //SSE4.1
+#   include <nmmintrin.h> //SSE4.2
 #endif
 
 #include <math.h>
@@ -67,32 +67,38 @@
 //***************  functions and data attributes, compiler dependent  *********************************
 //***********************************************************************************
 #ifdef __GNUC__
-#define _GCC_VERSION (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__)
-#define _NEON2SSE_ALIGN_16  __attribute__((aligned(16)))
-#define _NEON2SSE_INLINE extern inline __attribute__((__gnu_inline__, __always_inline__, __artificial__))
-#if _GCC_VERSION <  40500
-    #define _NEON2SSE_PERFORMANCE_WARNING(function, explanation)   __attribute__((deprecated)) function
-#else
-    #define _NEON2SSE_PERFORMANCE_WARNING(function, explanation)   __attribute__((deprecated(explanation))) function
-#endif
-#if defined(__x86_64__)
-    #define _NEON2SSE_64BIT  __x86_64__
-#endif
-#else
-#define _NEON2SSE_ALIGN_16  __declspec(align(16))
-#define _NEON2SSE_INLINE __inline
-#if defined(_MSC_VER)|| defined (__INTEL_COMPILER)  
-    #define _NEON2SSE_PERFORMANCE_WARNING(function, EXPLANATION) __declspec(deprecated(EXPLANATION)) function
-#if defined(_M_X64)
-        #define _NEON2SSE_64BIT  _M_X64
-#endif
+#   define _GCC_VERSION (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__)
+#   define _NEON2SSESTORAGE static
+#   define _NEON2SSE_ALIGN_16  __attribute__((aligned(16)))
+#   define _NEON2SSE_INLINE _NEON2SSESTORAGE inline __attribute__((__gnu_inline__, __always_inline__, __artificial__))
+#   ifndef NEON2SSE_DISABLE_PERFORMANCE_WARNING
+#       if _GCC_VERSION <  40500
+#           define _NEON2SSE_PERFORMANCE_WARNING(function, explanation)   __attribute__((deprecated)) function
+#       else
+#           define _NEON2SSE_PERFORMANCE_WARNING(function, explanation)   __attribute__((deprecated(explanation))) function
+#       endif
+#   else
+#       define _NEON2SSE_PERFORMANCE_WARNING(function, explanation)  function
+#   endif
+#   if defined(__x86_64__)
+#       define _NEON2SSE_64BIT  __x86_64__
+#   endif
 #else
-    #define _NEON2SSE_PERFORMANCE_WARNING(function, explanation)  function
-#endif
+#   define _NEON2SSESTORAGE static
+#   define _NEON2SSE_ALIGN_16  __declspec(align(16))
+#   define _NEON2SSE_INLINE _NEON2SSESTORAGE __inline
+#   if (defined(_MSC_VER) || defined (__INTEL_COMPILER)) && !defined(NEON2SSE_DISABLE_PERFORMANCE_WARNING)
+#       define _NEON2SSE_PERFORMANCE_WARNING(function, EXPLANATION) __declspec(deprecated(EXPLANATION)) function
+#       if defined(_M_X64)
+#           define _NEON2SSE_64BIT  _M_X64
+#       endif
+#   else
+#       define _NEON2SSE_PERFORMANCE_WARNING(function, explanation)  function
+#   endif
 #endif
 
 #if defined  (_NEON2SSE_64BIT) && defined (USE_SSE4)
-    #define _NEON2SSE_64BIT_SSE4
+#   define _NEON2SSE_64BIT_SSE4
 #endif
 
 /*********************************************************************************************************************/
@@ -118,8 +124,8 @@
     typedef signed long long int64_t;
     typedef unsigned long long uint64_t;
 #else
-#include <stdint.h>
-#include <limits.h>
+#   include <stdint.h>
+#   include <limits.h>
 #endif
 
 typedef union   __m64_128 {
@@ -166,11 +172,11 @@ typedef __m128i poly8x16_t;
 typedef __m128i poly16x8_t;
 
 #if defined(_MSC_VER)
-    #define SINT_MIN     (-2147483647 - 1) /* min signed int value */
-    #define SINT_MAX       2147483647 /* max signed int value */
+#   define SINT_MIN     (-2147483647 - 1) /* min signed int value */
+#   define SINT_MAX       2147483647 /* max signed int value */
 #else
-    #define SINT_MIN     INT_MIN /* min signed int value */
-    #define SINT_MAX     INT_MAX /* max signed int value */
+#   define SINT_MIN     INT_MIN /* min signed int value */
+#   define SINT_MAX     INT_MAX /* max signed int value */
 #endif
 
 typedef   float float32_t;
@@ -397,15 +403,14 @@ typedef  float16x8x3_t float16x4x3_t;
 #define _M128(a) _mm_castsi128_ps(a)
 //here the most performance effective implementation is compiler and 32/64 bits build dependent
 #if defined (_NEON2SSE_64BIT) || (defined (__INTEL_COMPILER) && (__INTEL_COMPILER  >= 1500) )
-
-        #define _pM128i(a) _mm_cvtsi64_si128(*(int64_t*)(&(a)))
-        #define _M64(out, inp) out.m64_i64[0] = _mm_cvtsi128_si64 (inp);
-        #define _M64f(out, inp) out.m64_i64[0] = _mm_cvtsi128_si64 (_M128i(inp));
-#else  
+#   define _pM128i(a) _mm_cvtsi64_si128(*(int64_t*)(&(a)))
+#   define _M64(out, inp) out.m64_i64[0] = _mm_cvtsi128_si64 (inp);
+#   define _M64f(out, inp) out.m64_i64[0] = _mm_cvtsi128_si64 (_M128i(inp));
+#else
    //for 32bit gcc and Microsoft compilers builds
-    #define _pM128i(a) _mm_loadl_epi64((__m128i*)&(a))
-    #define _M64(out, inp)  _mm_storel_epi64 ((__m128i*)&(out), inp)
-    #define _M64f(out, inp)  _mm_storel_epi64 ((__m128i*)&(out), _M128i(inp))
+#   define _pM128i(a) _mm_loadl_epi64((__m128i*)&(a))
+#   define _M64(out, inp)  _mm_storel_epi64 ((__m128i*)&(out), inp)
+#   define _M64f(out, inp)  _mm_storel_epi64 ((__m128i*)&(out), _M128i(inp))
 #endif
 #define _pM128(a) _mm_castsi128_ps(_pM128i(a))
 
@@ -428,8 +433,8 @@ typedef  float16x8x3_t float16x4x3_t;
 //~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 //&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&& mask constants used in porting &&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
-_NEON2SSE_ALIGN_16 static int8_t mask8_16_even_odd[16] = { 0, 2, 4, 6, 8, 10, 12, 14, 1, 3, 5, 7, 9, 11, 13, 15 };
-_NEON2SSE_ALIGN_16 static int8_t mask8_32_even_odd[16] = { 0,1, 4,5, 8,9, 12,13, 2,3, 6,7, 10,11, 14,15};
+_NEON2SSE_ALIGN_16 static const int8_t mask8_16_even_odd[16] = { 0, 2, 4, 6, 8, 10, 12, 14, 1, 3, 5, 7,  9, 11, 13, 15 };
+_NEON2SSE_ALIGN_16 static const int8_t mask8_32_even_odd[16] = { 0, 1, 4, 5, 8,  9, 12, 13, 2, 3, 6, 7, 10, 11, 14, 15 };
 //&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
 
 //*************************************************************************
@@ -437,1874 +442,1871 @@ _NEON2SSE_ALIGN_16 static int8_t mask8_32_even_odd[16] = { 0,1, 4,5, 8,9, 12,13,
 //*********  Functions declarations as declared in original arm_neon.h *****
 //*************************************************************************
 //Vector add: vadd -> Vr[i]:=Va[i]+Vb[i], Vr, Va, Vb have equal lane sizes.
-int8x8_t vadd_s8(int8x8_t a, int8x8_t b); // VADD.I8 d0,d0,d0
-int16x4_t vadd_s16(int16x4_t a, int16x4_t b); // VADD.I16 d0,d0,d0
-int32x2_t vadd_s32(int32x2_t a, int32x2_t b); // VADD.I32 d0,d0,d0
-int64x1_t vadd_s64(int64x1_t a, int64x1_t b); // VADD.I64 d0,d0,d0
-float32x2_t vadd_f32(float32x2_t a, float32x2_t b); // VADD.F32 d0,d0,d0
-uint8x8_t vadd_u8(uint8x8_t a, uint8x8_t b); // VADD.I8 d0,d0,d0
-uint16x4_t vadd_u16(uint16x4_t a, uint16x4_t b); // VADD.I16 d0,d0,d0
-uint32x2_t vadd_u32(uint32x2_t a, uint32x2_t b); // VADD.I32 d0,d0,d0
-uint64x1_t vadd_u64(uint64x1_t a, uint64x1_t b); // VADD.I64 d0,d0,d0
-int8x16_t vaddq_s8(int8x16_t a, int8x16_t b); // VADD.I8 q0,q0,q0
-int16x8_t vaddq_s16(int16x8_t a, int16x8_t b); // VADD.I16 q0,q0,q0
-int32x4_t vaddq_s32(int32x4_t a, int32x4_t b); // VADD.I32 q0,q0,q0
-int64x2_t vaddq_s64(int64x2_t a, int64x2_t b); // VADD.I64 q0,q0,q0
-float32x4_t vaddq_f32(float32x4_t a, float32x4_t b); // VADD.F32 q0,q0,q0
-uint8x16_t vaddq_u8(uint8x16_t a, uint8x16_t b); // VADD.I8 q0,q0,q0
-uint16x8_t vaddq_u16(uint16x8_t a, uint16x8_t b); // VADD.I16 q0,q0,q0
-uint32x4_t vaddq_u32(uint32x4_t a, uint32x4_t b); // VADD.I32 q0,q0,q0
-uint64x2_t vaddq_u64(uint64x2_t a, uint64x2_t b); // VADD.I64 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vadd_s8(int8x8_t a, int8x8_t b); // VADD.I8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vadd_s16(int16x4_t a, int16x4_t b); // VADD.I16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vadd_s32(int32x2_t a, int32x2_t b); // VADD.I32 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vadd_s64(int64x1_t a, int64x1_t b); // VADD.I64 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vadd_f32(float32x2_t a, float32x2_t b); // VADD.F32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vadd_u8(uint8x8_t a, uint8x8_t b); // VADD.I8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vadd_u16(uint16x4_t a, uint16x4_t b); // VADD.I16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vadd_u32(uint32x2_t a, uint32x2_t b); // VADD.I32 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vadd_u64(uint64x1_t a, uint64x1_t b); // VADD.I64 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vaddq_s8(int8x16_t a, int8x16_t b); // VADD.I8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vaddq_s16(int16x8_t a, int16x8_t b); // VADD.I16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vaddq_s32(int32x4_t a, int32x4_t b); // VADD.I32 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vaddq_s64(int64x2_t a, int64x2_t b); // VADD.I64 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vaddq_f32(float32x4_t a, float32x4_t b); // VADD.F32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vaddq_u8(uint8x16_t a, uint8x16_t b); // VADD.I8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vaddq_u16(uint16x8_t a, uint16x8_t b); // VADD.I16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vaddq_u32(uint32x4_t a, uint32x4_t b); // VADD.I32 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vaddq_u64(uint64x2_t a, uint64x2_t b); // VADD.I64 q0,q0,q0
 //Vector long add: vaddl -> Vr[i]:=Va[i]+Vb[i], Va, Vb have equal lane sizes, result is a 128 bit vector of lanes that are twice the width.
-int16x8_t vaddl_s8(int8x8_t a, int8x8_t b); // VADDL.S8 q0,d0,d0
-int32x4_t vaddl_s16(int16x4_t a, int16x4_t b); // VADDL.S16 q0,d0,d0
-int64x2_t vaddl_s32(int32x2_t a, int32x2_t b); // VADDL.S32 q0,d0,d0
-uint16x8_t vaddl_u8(uint8x8_t a, uint8x8_t b); // VADDL.U8 q0,d0,d0
-uint32x4_t vaddl_u16(uint16x4_t a, uint16x4_t b); // VADDL.U16 q0,d0,d0
-uint64x2_t vaddl_u32(uint32x2_t a, uint32x2_t b); // VADDL.U32 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vaddl_s8(int8x8_t a, int8x8_t b); // VADDL.S8 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vaddl_s16(int16x4_t a, int16x4_t b); // VADDL.S16 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vaddl_s32(int32x2_t a, int32x2_t b); // VADDL.S32 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vaddl_u8(uint8x8_t a, uint8x8_t b); // VADDL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vaddl_u16(uint16x4_t a, uint16x4_t b); // VADDL.U16 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vaddl_u32(uint32x2_t a, uint32x2_t b); // VADDL.U32 q0,d0,d0
 //Vector wide addw: vadd -> Vr[i]:=Va[i]+Vb[i]
-int16x8_t vaddw_s8(int16x8_t a, int8x8_t b); // VADDW.S8 q0,q0,d0
-int32x4_t vaddw_s16(int32x4_t a, int16x4_t b); // VADDW.S16 q0,q0,d0
-int64x2_t vaddw_s32(int64x2_t a, int32x2_t b); // VADDW.S32 q0,q0,d0
-uint16x8_t vaddw_u8(uint16x8_t a, uint8x8_t b); // VADDW.U8 q0,q0,d0
-uint32x4_t vaddw_u16(uint32x4_t a, uint16x4_t b); // VADDW.U16 q0,q0,d0
-uint64x2_t vaddw_u32(uint64x2_t a, uint32x2_t b); // VADDW.U32 q0,q0,d0
+_NEON2SSESTORAGE int16x8_t vaddw_s8(int16x8_t a, int8x8_t b); // VADDW.S8 q0,q0,d0
+_NEON2SSESTORAGE int32x4_t vaddw_s16(int32x4_t a, int16x4_t b); // VADDW.S16 q0,q0,d0
+_NEON2SSESTORAGE int64x2_t vaddw_s32(int64x2_t a, int32x2_t b); // VADDW.S32 q0,q0,d0
+_NEON2SSESTORAGE uint16x8_t vaddw_u8(uint16x8_t a, uint8x8_t b); // VADDW.U8 q0,q0,d0
+_NEON2SSESTORAGE uint32x4_t vaddw_u16(uint32x4_t a, uint16x4_t b); // VADDW.U16 q0,q0,d0
+_NEON2SSESTORAGE uint64x2_t vaddw_u32(uint64x2_t a, uint32x2_t b); // VADDW.U32 q0,q0,d0
 //Vector halving add: vhadd -> Vr[i]:=(Va[i]+Vb[i])>>1
-int8x8_t vhadd_s8(int8x8_t a, int8x8_t b); // VHADD.S8 d0,d0,d0
-int16x4_t vhadd_s16(int16x4_t a, int16x4_t b); // VHADD.S16 d0,d0,d0
-int32x2_t vhadd_s32(int32x2_t a, int32x2_t b); // VHADD.S32 d0,d0,d0
-uint8x8_t vhadd_u8(uint8x8_t a, uint8x8_t b); // VHADD.U8 d0,d0,d0
-uint16x4_t vhadd_u16(uint16x4_t a, uint16x4_t b); // VHADD.U16 d0,d0,d0
-uint32x2_t vhadd_u32(uint32x2_t a, uint32x2_t b); // VHADD.U32 d0,d0,d0
-int8x16_t vhaddq_s8(int8x16_t a, int8x16_t b); // VHADD.S8 q0,q0,q0
-int16x8_t vhaddq_s16(int16x8_t a, int16x8_t b); // VHADD.S16 q0,q0,q0
-int32x4_t vhaddq_s32(int32x4_t a, int32x4_t b); // VHADD.S32 q0,q0,q0
-uint8x16_t vhaddq_u8(uint8x16_t a, uint8x16_t b); // VHADD.U8 q0,q0,q0
-uint16x8_t vhaddq_u16(uint16x8_t a, uint16x8_t b); // VHADD.U16 q0,q0,q0
-uint32x4_t vhaddq_u32(uint32x4_t a, uint32x4_t b); // VHADD.U32 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vhadd_s8(int8x8_t a, int8x8_t b); // VHADD.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vhadd_s16(int16x4_t a, int16x4_t b); // VHADD.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vhadd_s32(int32x2_t a, int32x2_t b); // VHADD.S32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vhadd_u8(uint8x8_t a, uint8x8_t b); // VHADD.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vhadd_u16(uint16x4_t a, uint16x4_t b); // VHADD.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vhadd_u32(uint32x2_t a, uint32x2_t b); // VHADD.U32 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vhaddq_s8(int8x16_t a, int8x16_t b); // VHADD.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vhaddq_s16(int16x8_t a, int16x8_t b); // VHADD.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vhaddq_s32(int32x4_t a, int32x4_t b); // VHADD.S32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vhaddq_u8(uint8x16_t a, uint8x16_t b); // VHADD.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vhaddq_u16(uint16x8_t a, uint16x8_t b); // VHADD.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vhaddq_u32(uint32x4_t a, uint32x4_t b); // VHADD.U32 q0,q0,q0
 //Vector rounding halving add: vrhadd -> Vr[i]:=(Va[i]+Vb[i]+1)>>1
-int8x8_t vrhadd_s8(int8x8_t a, int8x8_t b); // VRHADD.S8 d0,d0,d0
-int16x4_t vrhadd_s16(int16x4_t a, int16x4_t b); // VRHADD.S16 d0,d0,d0
-int32x2_t vrhadd_s32(int32x2_t a, int32x2_t b); // VRHADD.S32 d0,d0,d0
-uint8x8_t vrhadd_u8(uint8x8_t a, uint8x8_t b); // VRHADD.U8 d0,d0,d0
-uint16x4_t vrhadd_u16(uint16x4_t a, uint16x4_t b); // VRHADD.U16 d0,d0,d0
-uint32x2_t vrhadd_u32(uint32x2_t a, uint32x2_t b); // VRHADD.U32 d0,d0,d0
-int8x16_t vrhaddq_s8(int8x16_t a, int8x16_t b); // VRHADD.S8 q0,q0,q0
-int16x8_t vrhaddq_s16(int16x8_t a, int16x8_t b); // VRHADD.S16 q0,q0,q0
-int32x4_t vrhaddq_s32(int32x4_t a, int32x4_t b); // VRHADD.S32 q0,q0,q0
-uint8x16_t vrhaddq_u8(uint8x16_t a, uint8x16_t b); // VRHADD.U8 q0,q0,q0
-uint16x8_t vrhaddq_u16(uint16x8_t a, uint16x8_t b); // VRHADD.U16 q0,q0,q0
-uint32x4_t vrhaddq_u32(uint32x4_t a, uint32x4_t b); // VRHADD.U32 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vrhadd_s8(int8x8_t a, int8x8_t b); // VRHADD.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vrhadd_s16(int16x4_t a, int16x4_t b); // VRHADD.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vrhadd_s32(int32x2_t a, int32x2_t b); // VRHADD.S32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vrhadd_u8(uint8x8_t a, uint8x8_t b); // VRHADD.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vrhadd_u16(uint16x4_t a, uint16x4_t b); // VRHADD.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vrhadd_u32(uint32x2_t a, uint32x2_t b); // VRHADD.U32 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vrhaddq_s8(int8x16_t a, int8x16_t b); // VRHADD.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vrhaddq_s16(int16x8_t a, int16x8_t b); // VRHADD.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vrhaddq_s32(int32x4_t a, int32x4_t b); // VRHADD.S32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vrhaddq_u8(uint8x16_t a, uint8x16_t b); // VRHADD.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vrhaddq_u16(uint16x8_t a, uint16x8_t b); // VRHADD.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vrhaddq_u32(uint32x4_t a, uint32x4_t b); // VRHADD.U32 q0,q0,q0
 //Vector saturating add: vqadd -> Vr[i]:=sat<size>(Va[i]+Vb[i])
-int8x8_t vqadd_s8(int8x8_t a, int8x8_t b); // VQADD.S8 d0,d0,d0
-int16x4_t vqadd_s16(int16x4_t a, int16x4_t b); // VQADD.S16 d0,d0,d0
-int32x2_t vqadd_s32(int32x2_t a, int32x2_t b); // VQADD.S32 d0,d0,d0
-int64x1_t vqadd_s64(int64x1_t a, int64x1_t b); // VQADD.S64 d0,d0,d0
-uint8x8_t vqadd_u8(uint8x8_t a, uint8x8_t b); // VQADD.U8 d0,d0,d0
-uint16x4_t vqadd_u16(uint16x4_t a, uint16x4_t b); // VQADD.U16 d0,d0,d0
-uint32x2_t vqadd_u32(uint32x2_t a, uint32x2_t b); // VQADD.U32 d0,d0,d0
-uint64x1_t vqadd_u64(uint64x1_t a, uint64x1_t b); // VQADD.U64 d0,d0,d0
-int8x16_t vqaddq_s8(int8x16_t a, int8x16_t b); // VQADD.S8 q0,q0,q0
-int16x8_t vqaddq_s16(int16x8_t a, int16x8_t b); // VQADD.S16 q0,q0,q0
-int32x4_t vqaddq_s32(int32x4_t a, int32x4_t b); // VQADD.S32 q0,q0,q0
-int64x2_t vqaddq_s64(int64x2_t a, int64x2_t b); // VQADD.S64 q0,q0,q0
-uint8x16_t vqaddq_u8(uint8x16_t a, uint8x16_t b); // VQADD.U8 q0,q0,q0
-uint16x8_t vqaddq_u16(uint16x8_t a, uint16x8_t b); // VQADD.U16 q0,q0,q0
-uint32x4_t vqaddq_u32(uint32x4_t a, uint32x4_t b); // VQADD.U32 q0,q0,q0
-uint64x2_t vqaddq_u64(uint64x2_t a, uint64x2_t b); // VQADD.U64 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vqadd_s8(int8x8_t a, int8x8_t b); // VQADD.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vqadd_s16(int16x4_t a, int16x4_t b); // VQADD.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqadd_s32(int32x2_t a, int32x2_t b); // VQADD.S32 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vqadd_s64(int64x1_t a, int64x1_t b); // VQADD.S64 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vqadd_u8(uint8x8_t a, uint8x8_t b); // VQADD.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vqadd_u16(uint16x4_t a, uint16x4_t b); // VQADD.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vqadd_u32(uint32x2_t a, uint32x2_t b); // VQADD.U32 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vqadd_u64(uint64x1_t a, uint64x1_t b); // VQADD.U64 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vqaddq_s8(int8x16_t a, int8x16_t b); // VQADD.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vqaddq_s16(int16x8_t a, int16x8_t b); // VQADD.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vqaddq_s32(int32x4_t a, int32x4_t b); // VQADD.S32 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vqaddq_s64(int64x2_t a, int64x2_t b); // VQADD.S64 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vqaddq_u8(uint8x16_t a, uint8x16_t b); // VQADD.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vqaddq_u16(uint16x8_t a, uint16x8_t b); // VQADD.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vqaddq_u32(uint32x4_t a, uint32x4_t b); // VQADD.U32 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vqaddq_u64(uint64x2_t a, uint64x2_t b); // VQADD.U64 q0,q0,q0
 //Vector add high half: vaddhn-> Vr[i]:=Va[i]+Vb[i]
-int8x8_t vaddhn_s16(int16x8_t a, int16x8_t b); // VADDHN.I16 d0,q0,q0
-int16x4_t vaddhn_s32(int32x4_t a, int32x4_t b); // VADDHN.I32 d0,q0,q0
-int32x2_t vaddhn_s64(int64x2_t a, int64x2_t b); // VADDHN.I64 d0,q0,q0
-uint8x8_t vaddhn_u16(uint16x8_t a, uint16x8_t b); // VADDHN.I16 d0,q0,q0
-uint16x4_t vaddhn_u32(uint32x4_t a, uint32x4_t b); // VADDHN.I32 d0,q0,q0
-uint32x2_t vaddhn_u64(uint64x2_t a, uint64x2_t b); // VADDHN.I64 d0,q0,q0
+_NEON2SSESTORAGE int8x8_t vaddhn_s16(int16x8_t a, int16x8_t b); // VADDHN.I16 d0,q0,q0
+_NEON2SSESTORAGE int16x4_t vaddhn_s32(int32x4_t a, int32x4_t b); // VADDHN.I32 d0,q0,q0
+_NEON2SSESTORAGE int32x2_t vaddhn_s64(int64x2_t a, int64x2_t b); // VADDHN.I64 d0,q0,q0
+_NEON2SSESTORAGE uint8x8_t vaddhn_u16(uint16x8_t a, uint16x8_t b); // VADDHN.I16 d0,q0,q0
+_NEON2SSESTORAGE uint16x4_t vaddhn_u32(uint32x4_t a, uint32x4_t b); // VADDHN.I32 d0,q0,q0
+_NEON2SSESTORAGE uint32x2_t vaddhn_u64(uint64x2_t a, uint64x2_t b); // VADDHN.I64 d0,q0,q0
 //Vector rounding add high half: vraddhn
-int8x8_t vraddhn_s16(int16x8_t a, int16x8_t b); // VRADDHN.I16 d0,q0,q0
-int16x4_t vraddhn_s32(int32x4_t a, int32x4_t b); // VRADDHN.I32 d0,q0,q0
-int32x2_t vraddhn_s64(int64x2_t a, int64x2_t b); // VRADDHN.I64 d0,q0,q0
-uint8x8_t vraddhn_u16(uint16x8_t a, uint16x8_t b); // VRADDHN.I16 d0,q0,q0
-uint16x4_t vraddhn_u32(uint32x4_t a, uint32x4_t b); // VRADDHN.I32 d0,q0,q0
-uint32x2_t vraddhn_u64(uint64x2_t a, uint64x2_t b); // VRADDHN.I64 d0,q0,q0
+_NEON2SSESTORAGE int8x8_t vraddhn_s16(int16x8_t a, int16x8_t b); // VRADDHN.I16 d0,q0,q0
+_NEON2SSESTORAGE int16x4_t vraddhn_s32(int32x4_t a, int32x4_t b); // VRADDHN.I32 d0,q0,q0
+_NEON2SSESTORAGE int32x2_t vraddhn_s64(int64x2_t a, int64x2_t b); // VRADDHN.I64 d0,q0,q0
+_NEON2SSESTORAGE uint8x8_t vraddhn_u16(uint16x8_t a, uint16x8_t b); // VRADDHN.I16 d0,q0,q0
+_NEON2SSESTORAGE uint16x4_t vraddhn_u32(uint32x4_t a, uint32x4_t b); // VRADDHN.I32 d0,q0,q0
+_NEON2SSESTORAGE uint32x2_t vraddhn_u64(uint64x2_t a, uint64x2_t b); // VRADDHN.I64 d0,q0,q0
 //Multiplication
 //Vector multiply: vmul -> Vr[i] := Va[i] * Vb[i]
-int8x8_t vmul_s8(int8x8_t a, int8x8_t b); // VMUL.I8 d0,d0,d0
-int16x4_t vmul_s16(int16x4_t a, int16x4_t b); // VMUL.I16 d0,d0,d0
-int32x2_t vmul_s32(int32x2_t a, int32x2_t b); // VMUL.I32 d0,d0,d0
-float32x2_t vmul_f32(float32x2_t a, float32x2_t b); // VMUL.F32 d0,d0,d0
-uint8x8_t vmul_u8(uint8x8_t a, uint8x8_t b); // VMUL.I8 d0,d0,d0
-uint16x4_t vmul_u16(uint16x4_t a, uint16x4_t b); // VMUL.I16 d0,d0,d0
-uint32x2_t vmul_u32(uint32x2_t a, uint32x2_t b); // VMUL.I32 d0,d0,d0
-poly8x8_t vmul_p8(poly8x8_t a, poly8x8_t b); // VMUL.P8 d0,d0,d0
-int8x16_t vmulq_s8(int8x16_t a, int8x16_t b); // VMUL.I8 q0,q0,q0
-int16x8_t vmulq_s16(int16x8_t a, int16x8_t b); // VMUL.I16 q0,q0,q0
-int32x4_t vmulq_s32(int32x4_t a, int32x4_t b); // VMUL.I32 q0,q0,q0
-float32x4_t vmulq_f32(float32x4_t a, float32x4_t b); // VMUL.F32 q0,q0,q0
-uint8x16_t vmulq_u8(uint8x16_t a, uint8x16_t b); // VMUL.I8 q0,q0,q0
-uint16x8_t vmulq_u16(uint16x8_t a, uint16x8_t b); // VMUL.I16 q0,q0,q0
-uint32x4_t vmulq_u32(uint32x4_t a, uint32x4_t b); // VMUL.I32 q0,q0,q0
-poly8x16_t vmulq_p8(poly8x16_t a, poly8x16_t b); // VMUL.P8 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vmul_s8(int8x8_t a, int8x8_t b); // VMUL.I8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vmul_s16(int16x4_t a, int16x4_t b); // VMUL.I16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vmul_s32(int32x2_t a, int32x2_t b); // VMUL.I32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vmul_f32(float32x2_t a, float32x2_t b); // VMUL.F32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vmul_u8(uint8x8_t a, uint8x8_t b); // VMUL.I8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vmul_u16(uint16x4_t a, uint16x4_t b); // VMUL.I16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vmul_u32(uint32x2_t a, uint32x2_t b); // VMUL.I32 d0,d0,d0
+_NEON2SSESTORAGE poly8x8_t vmul_p8(poly8x8_t a, poly8x8_t b); // VMUL.P8 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vmulq_s8(int8x16_t a, int8x16_t b); // VMUL.I8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vmulq_s16(int16x8_t a, int16x8_t b); // VMUL.I16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vmulq_s32(int32x4_t a, int32x4_t b); // VMUL.I32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vmulq_f32(float32x4_t a, float32x4_t b); // VMUL.F32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vmulq_u8(uint8x16_t a, uint8x16_t b); // VMUL.I8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vmulq_u16(uint16x8_t a, uint16x8_t b); // VMUL.I16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vmulq_u32(uint32x4_t a, uint32x4_t b); // VMUL.I32 q0,q0,q0
+_NEON2SSESTORAGE poly8x16_t vmulq_p8(poly8x16_t a, poly8x16_t b); // VMUL.P8 q0,q0,q0
 //multiply lane
-int16x4_t vmul_lane_s16 (int16x4_t a, int16x4_t b, __constrange(0,3) int c);
-int32x2_t vmul_lane_s32 (int32x2_t a, int32x2_t b, __constrange(0,1) int c);
-float32x2_t vmul_lane_f32 (float32x2_t a, float32x2_t b, __constrange(0,1) int c);
-uint16x4_t vmul_lane_u16 (uint16x4_t a, uint16x4_t b, __constrange(0,3) int c);
-uint32x2_t vmul_lane_u32 (uint32x2_t a, uint32x2_t b, __constrange(0,1) int c);
-int16x8_t vmulq_lane_s16 (int16x8_t a, int16x4_t b, __constrange(0,3) int c);
-int32x4_t vmulq_lane_s32 (int32x4_t a, int32x2_t b, __constrange(0,1) int c);
-float32x4_t vmulq_lane_f32 (float32x4_t a, float32x2_t b, __constrange(0,1) int c);
-uint16x8_t vmulq_lane_u16 (uint16x8_t a, uint16x4_t b, __constrange(0,3) int c);
-uint32x4_t vmulq_lane_u32 (uint32x4_t a, uint32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE int16x4_t vmul_lane_s16 (int16x4_t a, int16x4_t b, __constrange(0,3) int c);
+_NEON2SSESTORAGE int32x2_t vmul_lane_s32 (int32x2_t a, int32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE float32x2_t vmul_lane_f32 (float32x2_t a, float32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE uint16x4_t vmul_lane_u16 (uint16x4_t a, uint16x4_t b, __constrange(0,3) int c);
+_NEON2SSESTORAGE uint32x2_t vmul_lane_u32 (uint32x2_t a, uint32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE int16x8_t vmulq_lane_s16 (int16x8_t a, int16x4_t b, __constrange(0,3) int c);
+_NEON2SSESTORAGE int32x4_t vmulq_lane_s32 (int32x4_t a, int32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE float32x4_t vmulq_lane_f32 (float32x4_t a, float32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE uint16x8_t vmulq_lane_u16 (uint16x8_t a, uint16x4_t b, __constrange(0,3) int c);
+_NEON2SSESTORAGE uint32x4_t vmulq_lane_u32 (uint32x4_t a, uint32x2_t b, __constrange(0,1) int c);
 //Vector multiply accumulate: vmla -> Vr[i] := Va[i] + Vb[i] * Vc[i]
-int8x8_t vmla_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VMLA.I8 d0,d0,d0
-int16x4_t vmla_s16(int16x4_t a, int16x4_t b, int16x4_t c); // VMLA.I16 d0,d0,d0
-int32x2_t vmla_s32(int32x2_t a, int32x2_t b, int32x2_t c); // VMLA.I32 d0,d0,d0
-float32x2_t vmla_f32(float32x2_t a, float32x2_t b, float32x2_t c); // VMLA.F32 d0,d0,d0
-uint8x8_t vmla_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VMLA.I8 d0,d0,d0
-uint16x4_t vmla_u16(uint16x4_t a, uint16x4_t b, uint16x4_t c); // VMLA.I16 d0,d0,d0
-uint32x2_t vmla_u32(uint32x2_t a, uint32x2_t b, uint32x2_t c); // VMLA.I32 d0,d0,d0
-int8x16_t vmlaq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VMLA.I8 q0,q0,q0
-int16x8_t vmlaq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VMLA.I16 q0,q0,q0
-int32x4_t vmlaq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VMLA.I32 q0,q0,q0
-float32x4_t vmlaq_f32(float32x4_t a, float32x4_t b, float32x4_t c); // VMLA.F32 q0,q0,q0
-uint8x16_t vmlaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VMLA.I8 q0,q0,q0
-uint16x8_t vmlaq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VMLA.I16 q0,q0,q0
-uint32x4_t vmlaq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VMLA.I32 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vmla_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VMLA.I8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vmla_s16(int16x4_t a, int16x4_t b, int16x4_t c); // VMLA.I16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vmla_s32(int32x2_t a, int32x2_t b, int32x2_t c); // VMLA.I32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vmla_f32(float32x2_t a, float32x2_t b, float32x2_t c); // VMLA.F32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vmla_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VMLA.I8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vmla_u16(uint16x4_t a, uint16x4_t b, uint16x4_t c); // VMLA.I16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vmla_u32(uint32x2_t a, uint32x2_t b, uint32x2_t c); // VMLA.I32 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vmlaq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VMLA.I8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vmlaq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VMLA.I16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vmlaq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VMLA.I32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vmlaq_f32(float32x4_t a, float32x4_t b, float32x4_t c); // VMLA.F32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vmlaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VMLA.I8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vmlaq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VMLA.I16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vmlaq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VMLA.I32 q0,q0,q0
 //Vector multiply accumulate long: vmlal -> Vr[i] := Va[i] + Vb[i] * Vc[i]
-int16x8_t vmlal_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VMLAL.S8 q0,d0,d0
-int32x4_t vmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VMLAL.S16 q0,d0,d0
-int64x2_t vmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VMLAL.S32 q0,d0,d0
-uint16x8_t vmlal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VMLAL.U8 q0,d0,d0
-uint32x4_t vmlal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VMLAL.U16 q0,d0,d0
-uint64x2_t vmlal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VMLAL.U32 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vmlal_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VMLAL.S8 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VMLAL.S16 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VMLAL.S32 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vmlal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VMLAL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vmlal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VMLAL.U16 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vmlal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VMLAL.U32 q0,d0,d0
 //Vector multiply subtract: vmls -> Vr[i] := Va[i] - Vb[i] * Vc[i]
-int8x8_t vmls_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VMLS.I8 d0,d0,d0
-int16x4_t vmls_s16(int16x4_t a, int16x4_t b, int16x4_t c); // VMLS.I16 d0,d0,d0
-int32x2_t vmls_s32(int32x2_t a, int32x2_t b, int32x2_t c); // VMLS.I32 d0,d0,d0
-float32x2_t vmls_f32(float32x2_t a, float32x2_t b, float32x2_t c); // VMLS.F32 d0,d0,d0
-uint8x8_t vmls_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VMLS.I8 d0,d0,d0
-uint16x4_t vmls_u16(uint16x4_t a, uint16x4_t b, uint16x4_t c); // VMLS.I16 d0,d0,d0
-uint32x2_t vmls_u32(uint32x2_t a, uint32x2_t b, uint32x2_t c); // VMLS.I32 d0,d0,d0
-int8x16_t vmlsq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VMLS.I8 q0,q0,q0
-int16x8_t vmlsq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VMLS.I16 q0,q0,q0
-int32x4_t vmlsq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VMLS.I32 q0,q0,q0
-float32x4_t vmlsq_f32(float32x4_t a, float32x4_t b, float32x4_t c); // VMLS.F32 q0,q0,q0
-uint8x16_t vmlsq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VMLS.I8 q0,q0,q0
-uint16x8_t vmlsq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VMLS.I16 q0,q0,q0
-uint32x4_t vmlsq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VMLS.I32 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vmls_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VMLS.I8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vmls_s16(int16x4_t a, int16x4_t b, int16x4_t c); // VMLS.I16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vmls_s32(int32x2_t a, int32x2_t b, int32x2_t c); // VMLS.I32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vmls_f32(float32x2_t a, float32x2_t b, float32x2_t c); // VMLS.F32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vmls_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VMLS.I8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vmls_u16(uint16x4_t a, uint16x4_t b, uint16x4_t c); // VMLS.I16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vmls_u32(uint32x2_t a, uint32x2_t b, uint32x2_t c); // VMLS.I32 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vmlsq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VMLS.I8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vmlsq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VMLS.I16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vmlsq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VMLS.I32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vmlsq_f32(float32x4_t a, float32x4_t b, float32x4_t c); // VMLS.F32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vmlsq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VMLS.I8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vmlsq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VMLS.I16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vmlsq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VMLS.I32 q0,q0,q0
 //Vector multiply subtract long
-int16x8_t vmlsl_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VMLSL.S8 q0,d0,d0
-int32x4_t vmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VMLSL.S16 q0,d0,d0
-int64x2_t vmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VMLSL.S32 q0,d0,d0
-uint16x8_t vmlsl_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VMLSL.U8 q0,d0,d0
-uint32x4_t vmlsl_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VMLSL.U16 q0,d0,d0
-uint64x2_t vmlsl_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VMLSL.U32 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vmlsl_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VMLSL.S8 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VMLSL.S16 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VMLSL.S32 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vmlsl_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VMLSL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vmlsl_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VMLSL.U16 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vmlsl_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VMLSL.U32 q0,d0,d0
 //Vector saturating doubling multiply high
-int16x4_t vqdmulh_s16(int16x4_t a, int16x4_t b); // VQDMULH.S16 d0,d0,d0
-int32x2_t vqdmulh_s32(int32x2_t a, int32x2_t b); // VQDMULH.S32 d0,d0,d0
-int16x8_t vqdmulhq_s16(int16x8_t a, int16x8_t b); // VQDMULH.S16 q0,q0,q0
-int32x4_t vqdmulhq_s32(int32x4_t a, int32x4_t b); // VQDMULH.S32 q0,q0,q0
+_NEON2SSESTORAGE int16x4_t vqdmulh_s16(int16x4_t a, int16x4_t b); // VQDMULH.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqdmulh_s32(int32x2_t a, int32x2_t b); // VQDMULH.S32 d0,d0,d0
+_NEON2SSESTORAGE int16x8_t vqdmulhq_s16(int16x8_t a, int16x8_t b); // VQDMULH.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vqdmulhq_s32(int32x4_t a, int32x4_t b); // VQDMULH.S32 q0,q0,q0
 //Vector saturating rounding doubling multiply high
-int16x4_t vqrdmulh_s16(int16x4_t a, int16x4_t b); // VQRDMULH.S16 d0,d0,d0
-int32x2_t vqrdmulh_s32(int32x2_t a, int32x2_t b); // VQRDMULH.S32 d0,d0,d0
-int16x8_t vqrdmulhq_s16(int16x8_t a, int16x8_t b); // VQRDMULH.S16 q0,q0,q0
-int32x4_t vqrdmulhq_s32(int32x4_t a, int32x4_t b); // VQRDMULH.S32 q0,q0,q0
+_NEON2SSESTORAGE int16x4_t vqrdmulh_s16(int16x4_t a, int16x4_t b); // VQRDMULH.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqrdmulh_s32(int32x2_t a, int32x2_t b); // VQRDMULH.S32 d0,d0,d0
+_NEON2SSESTORAGE int16x8_t vqrdmulhq_s16(int16x8_t a, int16x8_t b); // VQRDMULH.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vqrdmulhq_s32(int32x4_t a, int32x4_t b); // VQRDMULH.S32 q0,q0,q0
 //Vector saturating doubling multiply accumulate long
-int32x4_t vqdmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VQDMLAL.S16 q0,d0,d0
-int64x2_t vqdmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VQDMLAL.S32 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vqdmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VQDMLAL.S16 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vqdmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VQDMLAL.S32 q0,d0,d0
 //Vector saturating doubling multiply subtract long
-int32x4_t vqdmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VQDMLSL.S16 q0,d0,d0
-int64x2_t vqdmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VQDMLSL.S32 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vqdmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VQDMLSL.S16 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vqdmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VQDMLSL.S32 q0,d0,d0
 //Vector long multiply
-int16x8_t vmull_s8(int8x8_t a, int8x8_t b); // VMULL.S8 q0,d0,d0
-int32x4_t vmull_s16(int16x4_t a, int16x4_t b); // VMULL.S16 q0,d0,d0
-int64x2_t vmull_s32(int32x2_t a, int32x2_t b); // VMULL.S32 q0,d0,d0
-uint16x8_t vmull_u8(uint8x8_t a, uint8x8_t b); // VMULL.U8 q0,d0,d0
-uint32x4_t vmull_u16(uint16x4_t a, uint16x4_t b); // VMULL.U16 q0,d0,d0
-uint64x2_t vmull_u32(uint32x2_t a, uint32x2_t b); // VMULL.U32 q0,d0,d0
-poly16x8_t vmull_p8(poly8x8_t a, poly8x8_t b); // VMULL.P8 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vmull_s8(int8x8_t a, int8x8_t b); // VMULL.S8 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vmull_s16(int16x4_t a, int16x4_t b); // VMULL.S16 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vmull_s32(int32x2_t a, int32x2_t b); // VMULL.S32 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vmull_u8(uint8x8_t a, uint8x8_t b); // VMULL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vmull_u16(uint16x4_t a, uint16x4_t b); // VMULL.U16 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vmull_u32(uint32x2_t a, uint32x2_t b); // VMULL.U32 q0,d0,d0
+_NEON2SSESTORAGE poly16x8_t vmull_p8(poly8x8_t a, poly8x8_t b); // VMULL.P8 q0,d0,d0
 //Vector saturating doubling long multiply
-int32x4_t vqdmull_s16(int16x4_t a, int16x4_t b); // VQDMULL.S16 q0,d0,d0
-int64x2_t vqdmull_s32(int32x2_t a, int32x2_t b); // VQDMULL.S32 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vqdmull_s16(int16x4_t a, int16x4_t b); // VQDMULL.S16 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vqdmull_s32(int32x2_t a, int32x2_t b); // VQDMULL.S32 q0,d0,d0
 //Subtraction
 //Vector subtract
-int8x8_t vsub_s8(int8x8_t a, int8x8_t b); // VSUB.I8 d0,d0,d0
-int16x4_t vsub_s16(int16x4_t a, int16x4_t b); // VSUB.I16 d0,d0,d0
-int32x2_t vsub_s32(int32x2_t a, int32x2_t b); // VSUB.I32 d0,d0,d0
-int64x1_t vsub_s64(int64x1_t a, int64x1_t b); // VSUB.I64 d0,d0,d0
-float32x2_t vsub_f32(float32x2_t a, float32x2_t b); // VSUB.F32 d0,d0,d0
-uint8x8_t vsub_u8(uint8x8_t a, uint8x8_t b); // VSUB.I8 d0,d0,d0
-uint16x4_t vsub_u16(uint16x4_t a, uint16x4_t b); // VSUB.I16 d0,d0,d0
-uint32x2_t vsub_u32(uint32x2_t a, uint32x2_t b); // VSUB.I32 d0,d0,d0
-uint64x1_t vsub_u64(uint64x1_t a, uint64x1_t b); // VSUB.I64 d0,d0,d0
-int8x16_t vsubq_s8(int8x16_t a, int8x16_t b); // VSUB.I8 q0,q0,q0
-int16x8_t vsubq_s16(int16x8_t a, int16x8_t b); // VSUB.I16 q0,q0,q0
-int32x4_t vsubq_s32(int32x4_t a, int32x4_t b); // VSUB.I32 q0,q0,q0
-int64x2_t vsubq_s64(int64x2_t a, int64x2_t b); // VSUB.I64 q0,q0,q0
-float32x4_t vsubq_f32(float32x4_t a, float32x4_t b); // VSUB.F32 q0,q0,q0
-uint8x16_t vsubq_u8(uint8x16_t a, uint8x16_t b); // VSUB.I8 q0,q0,q0
-uint16x8_t vsubq_u16(uint16x8_t a, uint16x8_t b); // VSUB.I16 q0,q0,q0
-uint32x4_t vsubq_u32(uint32x4_t a, uint32x4_t b); // VSUB.I32 q0,q0,q0
-uint64x2_t vsubq_u64(uint64x2_t a, uint64x2_t b); // VSUB.I64 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vsub_s8(int8x8_t a, int8x8_t b); // VSUB.I8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vsub_s16(int16x4_t a, int16x4_t b); // VSUB.I16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vsub_s32(int32x2_t a, int32x2_t b); // VSUB.I32 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vsub_s64(int64x1_t a, int64x1_t b); // VSUB.I64 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vsub_f32(float32x2_t a, float32x2_t b); // VSUB.F32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vsub_u8(uint8x8_t a, uint8x8_t b); // VSUB.I8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vsub_u16(uint16x4_t a, uint16x4_t b); // VSUB.I16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vsub_u32(uint32x2_t a, uint32x2_t b); // VSUB.I32 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vsub_u64(uint64x1_t a, uint64x1_t b); // VSUB.I64 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vsubq_s8(int8x16_t a, int8x16_t b); // VSUB.I8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vsubq_s16(int16x8_t a, int16x8_t b); // VSUB.I16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vsubq_s32(int32x4_t a, int32x4_t b); // VSUB.I32 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vsubq_s64(int64x2_t a, int64x2_t b); // VSUB.I64 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vsubq_f32(float32x4_t a, float32x4_t b); // VSUB.F32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vsubq_u8(uint8x16_t a, uint8x16_t b); // VSUB.I8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vsubq_u16(uint16x8_t a, uint16x8_t b); // VSUB.I16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vsubq_u32(uint32x4_t a, uint32x4_t b); // VSUB.I32 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vsubq_u64(uint64x2_t a, uint64x2_t b); // VSUB.I64 q0,q0,q0
 //Vector long subtract: vsub -> Vr[i]:=Va[i]+Vb[i]
-int16x8_t vsubl_s8(int8x8_t a, int8x8_t b); // VSUBL.S8 q0,d0,d0
-int32x4_t vsubl_s16(int16x4_t a, int16x4_t b); // VSUBL.S16 q0,d0,d0
-int64x2_t vsubl_s32(int32x2_t a, int32x2_t b); // VSUBL.S32 q0,d0,d0
-uint16x8_t vsubl_u8(uint8x8_t a, uint8x8_t b); // VSUBL.U8 q0,d0,d0
-uint32x4_t vsubl_u16(uint16x4_t a, uint16x4_t b); // VSUBL.U16 q0,d0,d0
-uint64x2_t vsubl_u32(uint32x2_t a, uint32x2_t b); // VSUBL.U32 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vsubl_s8(int8x8_t a, int8x8_t b); // VSUBL.S8 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vsubl_s16(int16x4_t a, int16x4_t b); // VSUBL.S16 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vsubl_s32(int32x2_t a, int32x2_t b); // VSUBL.S32 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vsubl_u8(uint8x8_t a, uint8x8_t b); // VSUBL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vsubl_u16(uint16x4_t a, uint16x4_t b); // VSUBL.U16 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vsubl_u32(uint32x2_t a, uint32x2_t b); // VSUBL.U32 q0,d0,d0
 //Vector wide subtract: vsub -> Vr[i]:=Va[i]+Vb[i]
-int16x8_t vsubw_s8(int16x8_t a, int8x8_t b); // VSUBW.S8 q0,q0,d0
-int32x4_t vsubw_s16(int32x4_t a, int16x4_t b); // VSUBW.S16 q0,q0,d0
-int64x2_t vsubw_s32(int64x2_t a, int32x2_t b); // VSUBW.S32 q0,q0,d0
-uint16x8_t vsubw_u8(uint16x8_t a, uint8x8_t b); // VSUBW.U8 q0,q0,d0
-uint32x4_t vsubw_u16(uint32x4_t a, uint16x4_t b); // VSUBW.U16 q0,q0,d0
-uint64x2_t vsubw_u32(uint64x2_t a, uint32x2_t b); // VSUBW.U32 q0,q0,d0
+_NEON2SSESTORAGE int16x8_t vsubw_s8(int16x8_t a, int8x8_t b); // VSUBW.S8 q0,q0,d0
+_NEON2SSESTORAGE int32x4_t vsubw_s16(int32x4_t a, int16x4_t b); // VSUBW.S16 q0,q0,d0
+_NEON2SSESTORAGE int64x2_t vsubw_s32(int64x2_t a, int32x2_t b); // VSUBW.S32 q0,q0,d0
+_NEON2SSESTORAGE uint16x8_t vsubw_u8(uint16x8_t a, uint8x8_t b); // VSUBW.U8 q0,q0,d0
+_NEON2SSESTORAGE uint32x4_t vsubw_u16(uint32x4_t a, uint16x4_t b); // VSUBW.U16 q0,q0,d0
+_NEON2SSESTORAGE uint64x2_t vsubw_u32(uint64x2_t a, uint32x2_t b); // VSUBW.U32 q0,q0,d0
 //Vector saturating subtract
-int8x8_t vqsub_s8(int8x8_t a, int8x8_t b); // VQSUB.S8 d0,d0,d0
-int16x4_t vqsub_s16(int16x4_t a, int16x4_t b); // VQSUB.S16 d0,d0,d0
-int32x2_t vqsub_s32(int32x2_t a, int32x2_t b); // VQSUB.S32 d0,d0,d0
-int64x1_t vqsub_s64(int64x1_t a, int64x1_t b); // VQSUB.S64 d0,d0,d0
-uint8x8_t vqsub_u8(uint8x8_t a, uint8x8_t b); // VQSUB.U8 d0,d0,d0
-uint16x4_t vqsub_u16(uint16x4_t a, uint16x4_t b); // VQSUB.U16 d0,d0,d0
-uint32x2_t vqsub_u32(uint32x2_t a, uint32x2_t b); // VQSUB.U32 d0,d0,d0
-uint64x1_t vqsub_u64(uint64x1_t a, uint64x1_t b); // VQSUB.U64 d0,d0,d0
-int8x16_t vqsubq_s8(int8x16_t a, int8x16_t b); // VQSUB.S8 q0,q0,q0
-int16x8_t vqsubq_s16(int16x8_t a, int16x8_t b); // VQSUB.S16 q0,q0,q0
-int32x4_t vqsubq_s32(int32x4_t a, int32x4_t b); // VQSUB.S32 q0,q0,q0
-int64x2_t vqsubq_s64(int64x2_t a, int64x2_t b); // VQSUB.S64 q0,q0,q0
-uint8x16_t vqsubq_u8(uint8x16_t a, uint8x16_t b); // VQSUB.U8 q0,q0,q0
-uint16x8_t vqsubq_u16(uint16x8_t a, uint16x8_t b); // VQSUB.U16 q0,q0,q0
-uint32x4_t vqsubq_u32(uint32x4_t a, uint32x4_t b); // VQSUB.U32 q0,q0,q0
-uint64x2_t vqsubq_u64(uint64x2_t a, uint64x2_t b); // VQSUB.U64 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vqsub_s8(int8x8_t a, int8x8_t b); // VQSUB.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vqsub_s16(int16x4_t a, int16x4_t b); // VQSUB.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqsub_s32(int32x2_t a, int32x2_t b); // VQSUB.S32 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vqsub_s64(int64x1_t a, int64x1_t b); // VQSUB.S64 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vqsub_u8(uint8x8_t a, uint8x8_t b); // VQSUB.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vqsub_u16(uint16x4_t a, uint16x4_t b); // VQSUB.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vqsub_u32(uint32x2_t a, uint32x2_t b); // VQSUB.U32 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vqsub_u64(uint64x1_t a, uint64x1_t b); // VQSUB.U64 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vqsubq_s8(int8x16_t a, int8x16_t b); // VQSUB.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vqsubq_s16(int16x8_t a, int16x8_t b); // VQSUB.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vqsubq_s32(int32x4_t a, int32x4_t b); // VQSUB.S32 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vqsubq_s64(int64x2_t a, int64x2_t b); // VQSUB.S64 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vqsubq_u8(uint8x16_t a, uint8x16_t b); // VQSUB.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vqsubq_u16(uint16x8_t a, uint16x8_t b); // VQSUB.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vqsubq_u32(uint32x4_t a, uint32x4_t b); // VQSUB.U32 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vqsubq_u64(uint64x2_t a, uint64x2_t b); // VQSUB.U64 q0,q0,q0
 //Vector halving subtract
-int8x8_t vhsub_s8(int8x8_t a, int8x8_t b); // VHSUB.S8 d0,d0,d0
-int16x4_t vhsub_s16(int16x4_t a, int16x4_t b); // VHSUB.S16 d0,d0,d0
-int32x2_t vhsub_s32(int32x2_t a, int32x2_t b); // VHSUB.S32 d0,d0,d0
-uint8x8_t vhsub_u8(uint8x8_t a, uint8x8_t b); // VHSUB.U8 d0,d0,d0
-uint16x4_t vhsub_u16(uint16x4_t a, uint16x4_t b); // VHSUB.U16 d0,d0,d0
-uint32x2_t vhsub_u32(uint32x2_t a, uint32x2_t b); // VHSUB.U32 d0,d0,d0
-int8x16_t vhsubq_s8(int8x16_t a, int8x16_t b); // VHSUB.S8 q0,q0,q0
-int16x8_t vhsubq_s16(int16x8_t a, int16x8_t b); // VHSUB.S16 q0,q0,q0
-int32x4_t vhsubq_s32(int32x4_t a, int32x4_t b); // VHSUB.S32 q0,q0,q0
-uint8x16_t vhsubq_u8(uint8x16_t a, uint8x16_t b); // VHSUB.U8 q0,q0,q0
-uint16x8_t vhsubq_u16(uint16x8_t a, uint16x8_t b); // VHSUB.U16 q0,q0,q0
-uint32x4_t vhsubq_u32(uint32x4_t a, uint32x4_t b); // VHSUB.U32 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vhsub_s8(int8x8_t a, int8x8_t b); // VHSUB.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vhsub_s16(int16x4_t a, int16x4_t b); // VHSUB.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vhsub_s32(int32x2_t a, int32x2_t b); // VHSUB.S32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vhsub_u8(uint8x8_t a, uint8x8_t b); // VHSUB.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vhsub_u16(uint16x4_t a, uint16x4_t b); // VHSUB.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vhsub_u32(uint32x2_t a, uint32x2_t b); // VHSUB.U32 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vhsubq_s8(int8x16_t a, int8x16_t b); // VHSUB.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vhsubq_s16(int16x8_t a, int16x8_t b); // VHSUB.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vhsubq_s32(int32x4_t a, int32x4_t b); // VHSUB.S32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vhsubq_u8(uint8x16_t a, uint8x16_t b); // VHSUB.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vhsubq_u16(uint16x8_t a, uint16x8_t b); // VHSUB.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vhsubq_u32(uint32x4_t a, uint32x4_t b); // VHSUB.U32 q0,q0,q0
 //Vector subtract high half
-int8x8_t vsubhn_s16(int16x8_t a, int16x8_t b); // VSUBHN.I16 d0,q0,q0
-int16x4_t vsubhn_s32(int32x4_t a, int32x4_t b); // VSUBHN.I32 d0,q0,q0
-int32x2_t vsubhn_s64(int64x2_t a, int64x2_t b); // VSUBHN.I64 d0,q0,q0
-uint8x8_t vsubhn_u16(uint16x8_t a, uint16x8_t b); // VSUBHN.I16 d0,q0,q0
-uint16x4_t vsubhn_u32(uint32x4_t a, uint32x4_t b); // VSUBHN.I32 d0,q0,q0
-uint32x2_t vsubhn_u64(uint64x2_t a, uint64x2_t b); // VSUBHN.I64 d0,q0,q0
+_NEON2SSESTORAGE int8x8_t vsubhn_s16(int16x8_t a, int16x8_t b); // VSUBHN.I16 d0,q0,q0
+_NEON2SSESTORAGE int16x4_t vsubhn_s32(int32x4_t a, int32x4_t b); // VSUBHN.I32 d0,q0,q0
+_NEON2SSESTORAGE int32x2_t vsubhn_s64(int64x2_t a, int64x2_t b); // VSUBHN.I64 d0,q0,q0
+_NEON2SSESTORAGE uint8x8_t vsubhn_u16(uint16x8_t a, uint16x8_t b); // VSUBHN.I16 d0,q0,q0
+_NEON2SSESTORAGE uint16x4_t vsubhn_u32(uint32x4_t a, uint32x4_t b); // VSUBHN.I32 d0,q0,q0
+_NEON2SSESTORAGE uint32x2_t vsubhn_u64(uint64x2_t a, uint64x2_t b); // VSUBHN.I64 d0,q0,q0
 //Vector rounding subtract high half
-int8x8_t vrsubhn_s16(int16x8_t a, int16x8_t b); // VRSUBHN.I16 d0,q0,q0
-int16x4_t vrsubhn_s32(int32x4_t a, int32x4_t b); // VRSUBHN.I32 d0,q0,q0
-int32x2_t vrsubhn_s64(int64x2_t a, int64x2_t b); // VRSUBHN.I64 d0,q0,q0
-uint8x8_t vrsubhn_u16(uint16x8_t a, uint16x8_t b); // VRSUBHN.I16 d0,q0,q0
-uint16x4_t vrsubhn_u32(uint32x4_t a, uint32x4_t b); // VRSUBHN.I32 d0,q0,q0
-uint32x2_t vrsubhn_u64(uint64x2_t a, uint64x2_t b); // VRSUBHN.I64 d0,q0,q0
+_NEON2SSESTORAGE int8x8_t vrsubhn_s16(int16x8_t a, int16x8_t b); // VRSUBHN.I16 d0,q0,q0
+_NEON2SSESTORAGE int16x4_t vrsubhn_s32(int32x4_t a, int32x4_t b); // VRSUBHN.I32 d0,q0,q0
+_NEON2SSESTORAGE int32x2_t vrsubhn_s64(int64x2_t a, int64x2_t b); // VRSUBHN.I64 d0,q0,q0
+_NEON2SSESTORAGE uint8x8_t vrsubhn_u16(uint16x8_t a, uint16x8_t b); // VRSUBHN.I16 d0,q0,q0
+_NEON2SSESTORAGE uint16x4_t vrsubhn_u32(uint32x4_t a, uint32x4_t b); // VRSUBHN.I32 d0,q0,q0
+_NEON2SSESTORAGE uint32x2_t vrsubhn_u64(uint64x2_t a, uint64x2_t b); // VRSUBHN.I64 d0,q0,q0
 //Comparison
 //Vector compare equal
-uint8x8_t vceq_s8(int8x8_t a, int8x8_t b); // VCEQ.I8 d0, d0, d0
-uint16x4_t vceq_s16(int16x4_t a, int16x4_t b); // VCEQ.I16 d0, d0, d0
-uint32x2_t vceq_s32(int32x2_t a, int32x2_t b); // VCEQ.I32 d0, d0, d0
-uint32x2_t vceq_f32(float32x2_t a, float32x2_t b); // VCEQ.F32 d0, d0, d0
-uint8x8_t vceq_u8(uint8x8_t a, uint8x8_t b); // VCEQ.I8 d0, d0, d0
-uint16x4_t vceq_u16(uint16x4_t a, uint16x4_t b); // VCEQ.I16 d0, d0, d0
-uint32x2_t vceq_u32(uint32x2_t a, uint32x2_t b); // VCEQ.I32 d0, d0, d0
-uint8x8_t vceq_p8(poly8x8_t a, poly8x8_t b); // VCEQ.I8 d0, d0, d0
-uint8x16_t vceqq_s8(int8x16_t a, int8x16_t b); // VCEQ.I8 q0, q0, q0
-uint16x8_t vceqq_s16(int16x8_t a, int16x8_t b); // VCEQ.I16 q0, q0, q0
-uint32x4_t vceqq_s32(int32x4_t a, int32x4_t b); // VCEQ.I32 q0, q0, q0
-uint32x4_t vceqq_f32(float32x4_t a, float32x4_t b); // VCEQ.F32 q0, q0, q0
-uint8x16_t vceqq_u8(uint8x16_t a, uint8x16_t b); // VCEQ.I8 q0, q0, q0
-uint16x8_t vceqq_u16(uint16x8_t a, uint16x8_t b); // VCEQ.I16 q0, q0, q0
-uint32x4_t vceqq_u32(uint32x4_t a, uint32x4_t b); // VCEQ.I32 q0, q0, q0
-uint8x16_t vceqq_p8(poly8x16_t a, poly8x16_t b); // VCEQ.I8 q0, q0, q0
+_NEON2SSESTORAGE uint8x8_t vceq_s8(int8x8_t a, int8x8_t b); // VCEQ.I8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vceq_s16(int16x4_t a, int16x4_t b); // VCEQ.I16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vceq_s32(int32x2_t a, int32x2_t b); // VCEQ.I32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vceq_f32(float32x2_t a, float32x2_t b); // VCEQ.F32 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vceq_u8(uint8x8_t a, uint8x8_t b); // VCEQ.I8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vceq_u16(uint16x4_t a, uint16x4_t b); // VCEQ.I16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vceq_u32(uint32x2_t a, uint32x2_t b); // VCEQ.I32 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vceq_p8(poly8x8_t a, poly8x8_t b); // VCEQ.I8 d0, d0, d0
+_NEON2SSESTORAGE uint8x16_t vceqq_s8(int8x16_t a, int8x16_t b); // VCEQ.I8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vceqq_s16(int16x8_t a, int16x8_t b); // VCEQ.I16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vceqq_s32(int32x4_t a, int32x4_t b); // VCEQ.I32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vceqq_f32(float32x4_t a, float32x4_t b); // VCEQ.F32 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vceqq_u8(uint8x16_t a, uint8x16_t b); // VCEQ.I8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vceqq_u16(uint16x8_t a, uint16x8_t b); // VCEQ.I16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vceqq_u32(uint32x4_t a, uint32x4_t b); // VCEQ.I32 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vceqq_p8(poly8x16_t a, poly8x16_t b); // VCEQ.I8 q0, q0, q0
 //Vector compare greater-than or equal
-uint8x8_t vcge_s8(int8x8_t a, int8x8_t b); // VCGE.S8 d0, d0, d0
-uint16x4_t vcge_s16(int16x4_t a, int16x4_t b); // VCGE.S16 d0, d0, d0
-uint32x2_t vcge_s32(int32x2_t a, int32x2_t b); // VCGE.S32 d0, d0, d0
-uint32x2_t vcge_f32(float32x2_t a, float32x2_t b); // VCGE.F32 d0, d0, d0
-uint8x8_t vcge_u8(uint8x8_t a, uint8x8_t b); // VCGE.U8 d0, d0, d0
-uint16x4_t vcge_u16(uint16x4_t a, uint16x4_t b); // VCGE.U16 d0, d0, d0
-uint32x2_t vcge_u32(uint32x2_t a, uint32x2_t b); // VCGE.U32 d0, d0, d0
-uint8x16_t vcgeq_s8(int8x16_t a, int8x16_t b); // VCGE.S8 q0, q0, q0
-uint16x8_t vcgeq_s16(int16x8_t a, int16x8_t b); // VCGE.S16 q0, q0, q0
-uint32x4_t vcgeq_s32(int32x4_t a, int32x4_t b); // VCGE.S32 q0, q0, q0
-uint32x4_t vcgeq_f32(float32x4_t a, float32x4_t b); // VCGE.F32 q0, q0, q0
-uint8x16_t vcgeq_u8(uint8x16_t a, uint8x16_t b); // VCGE.U8 q0, q0, q0
-uint16x8_t vcgeq_u16(uint16x8_t a, uint16x8_t b); // VCGE.U16 q0, q0, q0
-uint32x4_t vcgeq_u32(uint32x4_t a, uint32x4_t b); // VCGE.U32 q0, q0, q0
+_NEON2SSESTORAGE uint8x8_t vcge_s8(int8x8_t a, int8x8_t b); // VCGE.S8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcge_s16(int16x4_t a, int16x4_t b); // VCGE.S16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcge_s32(int32x2_t a, int32x2_t b); // VCGE.S32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcge_f32(float32x2_t a, float32x2_t b); // VCGE.F32 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vcge_u8(uint8x8_t a, uint8x8_t b); // VCGE.U8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcge_u16(uint16x4_t a, uint16x4_t b); // VCGE.U16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcge_u32(uint32x2_t a, uint32x2_t b); // VCGE.U32 d0, d0, d0
+_NEON2SSESTORAGE uint8x16_t vcgeq_s8(int8x16_t a, int8x16_t b); // VCGE.S8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcgeq_s16(int16x8_t a, int16x8_t b); // VCGE.S16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcgeq_s32(int32x4_t a, int32x4_t b); // VCGE.S32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcgeq_f32(float32x4_t a, float32x4_t b); // VCGE.F32 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vcgeq_u8(uint8x16_t a, uint8x16_t b); // VCGE.U8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcgeq_u16(uint16x8_t a, uint16x8_t b); // VCGE.U16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcgeq_u32(uint32x4_t a, uint32x4_t b); // VCGE.U32 q0, q0, q0
 //Vector compare less-than or equal
-uint8x8_t vcle_s8(int8x8_t a, int8x8_t b); // VCGE.S8 d0, d0, d0
-uint16x4_t vcle_s16(int16x4_t a, int16x4_t b); // VCGE.S16 d0, d0, d0
-uint32x2_t vcle_s32(int32x2_t a, int32x2_t b); // VCGE.S32 d0, d0, d0
-uint32x2_t vcle_f32(float32x2_t a, float32x2_t b); // VCGE.F32 d0, d0, d0
-uint8x8_t vcle_u8(uint8x8_t a, uint8x8_t b); // VCGE.U8 d0, d0, d0
-uint16x4_t vcle_u16(uint16x4_t a, uint16x4_t b); // VCGE.U16 d0, d0, d0
-uint32x2_t vcle_u32(uint32x2_t a, uint32x2_t b); // VCGE.U32 d0, d0, d0
-uint8x16_t vcleq_s8(int8x16_t a, int8x16_t b); // VCGE.S8 q0, q0, q0
-uint16x8_t vcleq_s16(int16x8_t a, int16x8_t b); // VCGE.S16 q0, q0, q0
-uint32x4_t vcleq_s32(int32x4_t a, int32x4_t b); // VCGE.S32 q0, q0, q0
-uint32x4_t vcleq_f32(float32x4_t a, float32x4_t b); // VCGE.F32 q0, q0, q0
-uint8x16_t vcleq_u8(uint8x16_t a, uint8x16_t b); // VCGE.U8 q0, q0, q0
-uint16x8_t vcleq_u16(uint16x8_t a, uint16x8_t b); // VCGE.U16 q0, q0, q0
-uint32x4_t vcleq_u32(uint32x4_t a, uint32x4_t b); // VCGE.U32 q0, q0, q0
+_NEON2SSESTORAGE uint8x8_t vcle_s8(int8x8_t a, int8x8_t b); // VCGE.S8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcle_s16(int16x4_t a, int16x4_t b); // VCGE.S16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcle_s32(int32x2_t a, int32x2_t b); // VCGE.S32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcle_f32(float32x2_t a, float32x2_t b); // VCGE.F32 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vcle_u8(uint8x8_t a, uint8x8_t b); // VCGE.U8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcle_u16(uint16x4_t a, uint16x4_t b); // VCGE.U16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcle_u32(uint32x2_t a, uint32x2_t b); // VCGE.U32 d0, d0, d0
+_NEON2SSESTORAGE uint8x16_t vcleq_s8(int8x16_t a, int8x16_t b); // VCGE.S8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcleq_s16(int16x8_t a, int16x8_t b); // VCGE.S16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcleq_s32(int32x4_t a, int32x4_t b); // VCGE.S32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcleq_f32(float32x4_t a, float32x4_t b); // VCGE.F32 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vcleq_u8(uint8x16_t a, uint8x16_t b); // VCGE.U8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcleq_u16(uint16x8_t a, uint16x8_t b); // VCGE.U16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcleq_u32(uint32x4_t a, uint32x4_t b); // VCGE.U32 q0, q0, q0
 //Vector compare greater-than
-uint8x8_t vcgt_s8(int8x8_t a, int8x8_t b); // VCGT.S8 d0, d0, d0
-uint16x4_t vcgt_s16(int16x4_t a, int16x4_t b); // VCGT.S16 d0, d0, d0
-uint32x2_t vcgt_s32(int32x2_t a, int32x2_t b); // VCGT.S32 d0, d0, d0
-uint32x2_t vcgt_f32(float32x2_t a, float32x2_t b); // VCGT.F32 d0, d0, d0
-uint8x8_t vcgt_u8(uint8x8_t a, uint8x8_t b); // VCGT.U8 d0, d0, d0
-uint16x4_t vcgt_u16(uint16x4_t a, uint16x4_t b); // VCGT.U16 d0, d0, d0
-uint32x2_t vcgt_u32(uint32x2_t a, uint32x2_t b); // VCGT.U32 d0, d0, d0
-uint8x16_t vcgtq_s8(int8x16_t a, int8x16_t b); // VCGT.S8 q0, q0, q0
-uint16x8_t vcgtq_s16(int16x8_t a, int16x8_t b); // VCGT.S16 q0, q0, q0
-uint32x4_t vcgtq_s32(int32x4_t a, int32x4_t b); // VCGT.S32 q0, q0, q0
-uint32x4_t vcgtq_f32(float32x4_t a, float32x4_t b); // VCGT.F32 q0, q0, q0
-uint8x16_t vcgtq_u8(uint8x16_t a, uint8x16_t b); // VCGT.U8 q0, q0, q0
-uint16x8_t vcgtq_u16(uint16x8_t a, uint16x8_t b); // VCGT.U16 q0, q0, q0
-uint32x4_t vcgtq_u32(uint32x4_t a, uint32x4_t b); // VCGT.U32 q0, q0, q0
+_NEON2SSESTORAGE uint8x8_t vcgt_s8(int8x8_t a, int8x8_t b); // VCGT.S8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcgt_s16(int16x4_t a, int16x4_t b); // VCGT.S16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcgt_s32(int32x2_t a, int32x2_t b); // VCGT.S32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcgt_f32(float32x2_t a, float32x2_t b); // VCGT.F32 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vcgt_u8(uint8x8_t a, uint8x8_t b); // VCGT.U8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcgt_u16(uint16x4_t a, uint16x4_t b); // VCGT.U16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcgt_u32(uint32x2_t a, uint32x2_t b); // VCGT.U32 d0, d0, d0
+_NEON2SSESTORAGE uint8x16_t vcgtq_s8(int8x16_t a, int8x16_t b); // VCGT.S8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcgtq_s16(int16x8_t a, int16x8_t b); // VCGT.S16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcgtq_s32(int32x4_t a, int32x4_t b); // VCGT.S32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcgtq_f32(float32x4_t a, float32x4_t b); // VCGT.F32 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vcgtq_u8(uint8x16_t a, uint8x16_t b); // VCGT.U8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcgtq_u16(uint16x8_t a, uint16x8_t b); // VCGT.U16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcgtq_u32(uint32x4_t a, uint32x4_t b); // VCGT.U32 q0, q0, q0
 //Vector compare less-than
-uint8x8_t vclt_s8(int8x8_t a, int8x8_t b); // VCGT.S8 d0, d0, d0
-uint16x4_t vclt_s16(int16x4_t a, int16x4_t b); // VCGT.S16 d0, d0, d0
-uint32x2_t vclt_s32(int32x2_t a, int32x2_t b); // VCGT.S32 d0, d0, d0
-uint32x2_t vclt_f32(float32x2_t a, float32x2_t b); // VCGT.F32 d0, d0, d0
-uint8x8_t vclt_u8(uint8x8_t a, uint8x8_t b); // VCGT.U8 d0, d0, d0
-uint16x4_t vclt_u16(uint16x4_t a, uint16x4_t b); // VCGT.U16 d0, d0, d0
-uint32x2_t vclt_u32(uint32x2_t a, uint32x2_t b); // VCGT.U32 d0, d0, d0
-uint8x16_t vcltq_s8(int8x16_t a, int8x16_t b); // VCGT.S8 q0, q0, q0
-uint16x8_t vcltq_s16(int16x8_t a, int16x8_t b); // VCGT.S16 q0, q0, q0
-uint32x4_t vcltq_s32(int32x4_t a, int32x4_t b); // VCGT.S32 q0, q0, q0
-uint32x4_t vcltq_f32(float32x4_t a, float32x4_t b); // VCGT.F32 q0, q0, q0
-uint8x16_t vcltq_u8(uint8x16_t a, uint8x16_t b); // VCGT.U8 q0, q0, q0
-uint16x8_t vcltq_u16(uint16x8_t a, uint16x8_t b); // VCGT.U16 q0, q0, q0
-uint32x4_t vcltq_u32(uint32x4_t a, uint32x4_t b); // VCGT.U32 q0, q0, q0
+_NEON2SSESTORAGE uint8x8_t vclt_s8(int8x8_t a, int8x8_t b); // VCGT.S8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vclt_s16(int16x4_t a, int16x4_t b); // VCGT.S16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vclt_s32(int32x2_t a, int32x2_t b); // VCGT.S32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vclt_f32(float32x2_t a, float32x2_t b); // VCGT.F32 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vclt_u8(uint8x8_t a, uint8x8_t b); // VCGT.U8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vclt_u16(uint16x4_t a, uint16x4_t b); // VCGT.U16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vclt_u32(uint32x2_t a, uint32x2_t b); // VCGT.U32 d0, d0, d0
+_NEON2SSESTORAGE uint8x16_t vcltq_s8(int8x16_t a, int8x16_t b); // VCGT.S8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcltq_s16(int16x8_t a, int16x8_t b); // VCGT.S16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcltq_s32(int32x4_t a, int32x4_t b); // VCGT.S32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcltq_f32(float32x4_t a, float32x4_t b); // VCGT.F32 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vcltq_u8(uint8x16_t a, uint8x16_t b); // VCGT.U8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcltq_u16(uint16x8_t a, uint16x8_t b); // VCGT.U16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcltq_u32(uint32x4_t a, uint32x4_t b); // VCGT.U32 q0, q0, q0
 //Vector compare absolute greater-than or equal
-uint32x2_t vcage_f32(float32x2_t a, float32x2_t b); // VACGE.F32 d0, d0, d0
-uint32x4_t vcageq_f32(float32x4_t a, float32x4_t b); // VACGE.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x2_t vcage_f32(float32x2_t a, float32x2_t b); // VACGE.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x4_t vcageq_f32(float32x4_t a, float32x4_t b); // VACGE.F32 q0, q0, q0
 //Vector compare absolute less-than or equal
-uint32x2_t vcale_f32(float32x2_t a, float32x2_t b); // VACGE.F32 d0, d0, d0
-uint32x4_t vcaleq_f32(float32x4_t a, float32x4_t b); // VACGE.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x2_t vcale_f32(float32x2_t a, float32x2_t b); // VACGE.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x4_t vcaleq_f32(float32x4_t a, float32x4_t b); // VACGE.F32 q0, q0, q0
 //Vector compare absolute greater-than
-uint32x2_t vcagt_f32(float32x2_t a, float32x2_t b); // VACGT.F32 d0, d0, d0
-uint32x4_t vcagtq_f32(float32x4_t a, float32x4_t b); // VACGT.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x2_t vcagt_f32(float32x2_t a, float32x2_t b); // VACGT.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x4_t vcagtq_f32(float32x4_t a, float32x4_t b); // VACGT.F32 q0, q0, q0
 //Vector compare absolute less-than
-uint32x2_t vcalt_f32(float32x2_t a, float32x2_t b); // VACGT.F32 d0, d0, d0
-uint32x4_t vcaltq_f32(float32x4_t a, float32x4_t b); // VACGT.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x2_t vcalt_f32(float32x2_t a, float32x2_t b); // VACGT.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x4_t vcaltq_f32(float32x4_t a, float32x4_t b); // VACGT.F32 q0, q0, q0
 //Vector test bits
-uint8x8_t vtst_s8(int8x8_t a, int8x8_t b); // VTST.8 d0, d0, d0
-uint16x4_t vtst_s16(int16x4_t a, int16x4_t b); // VTST.16 d0, d0, d0
-uint32x2_t vtst_s32(int32x2_t a, int32x2_t b); // VTST.32 d0, d0, d0
-uint8x8_t vtst_u8(uint8x8_t a, uint8x8_t b); // VTST.8 d0, d0, d0
-uint16x4_t vtst_u16(uint16x4_t a, uint16x4_t b); // VTST.16 d0, d0, d0
-uint32x2_t vtst_u32(uint32x2_t a, uint32x2_t b); // VTST.32 d0, d0, d0
-uint8x8_t vtst_p8(poly8x8_t a, poly8x8_t b); // VTST.8 d0, d0, d0
-uint8x16_t vtstq_s8(int8x16_t a, int8x16_t b); // VTST.8 q0, q0, q0
-uint16x8_t vtstq_s16(int16x8_t a, int16x8_t b); // VTST.16 q0, q0, q0
-uint32x4_t vtstq_s32(int32x4_t a, int32x4_t b); // VTST.32 q0, q0, q0
-uint8x16_t vtstq_u8(uint8x16_t a, uint8x16_t b); // VTST.8 q0, q0, q0
-uint16x8_t vtstq_u16(uint16x8_t a, uint16x8_t b); // VTST.16 q0, q0, q0
-uint32x4_t vtstq_u32(uint32x4_t a, uint32x4_t b); // VTST.32 q0, q0, q0
-uint8x16_t vtstq_p8(poly8x16_t a, poly8x16_t b); // VTST.8 q0, q0, q0
+_NEON2SSESTORAGE uint8x8_t vtst_s8(int8x8_t a, int8x8_t b); // VTST.8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vtst_s16(int16x4_t a, int16x4_t b); // VTST.16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vtst_s32(int32x2_t a, int32x2_t b); // VTST.32 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vtst_u8(uint8x8_t a, uint8x8_t b); // VTST.8 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vtst_u16(uint16x4_t a, uint16x4_t b); // VTST.16 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vtst_u32(uint32x2_t a, uint32x2_t b); // VTST.32 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vtst_p8(poly8x8_t a, poly8x8_t b); // VTST.8 d0, d0, d0
+_NEON2SSESTORAGE uint8x16_t vtstq_s8(int8x16_t a, int8x16_t b); // VTST.8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vtstq_s16(int16x8_t a, int16x8_t b); // VTST.16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vtstq_s32(int32x4_t a, int32x4_t b); // VTST.32 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vtstq_u8(uint8x16_t a, uint8x16_t b); // VTST.8 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vtstq_u16(uint16x8_t a, uint16x8_t b); // VTST.16 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vtstq_u32(uint32x4_t a, uint32x4_t b); // VTST.32 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vtstq_p8(poly8x16_t a, poly8x16_t b); // VTST.8 q0, q0, q0
 //Absolute difference
 //Absolute difference between the arguments: Vr[i] = | Va[i] - Vb[i] |
-int8x8_t vabd_s8(int8x8_t a, int8x8_t b); // VABD.S8 d0,d0,d0
-int16x4_t vabd_s16(int16x4_t a, int16x4_t b); // VABD.S16 d0,d0,d0
-int32x2_t vabd_s32(int32x2_t a, int32x2_t b); // VABD.S32 d0,d0,d0
-uint8x8_t vabd_u8(uint8x8_t a, uint8x8_t b); // VABD.U8 d0,d0,d0
-uint16x4_t vabd_u16(uint16x4_t a, uint16x4_t b); // VABD.U16 d0,d0,d0
-uint32x2_t vabd_u32(uint32x2_t a, uint32x2_t b); // VABD.U32 d0,d0,d0
-float32x2_t vabd_f32(float32x2_t a, float32x2_t b); // VABD.F32 d0,d0,d0
-int8x16_t vabdq_s8(int8x16_t a, int8x16_t b); // VABD.S8 q0,q0,q0
-int16x8_t vabdq_s16(int16x8_t a, int16x8_t b); // VABD.S16 q0,q0,q0
-int32x4_t vabdq_s32(int32x4_t a, int32x4_t b); // VABD.S32 q0,q0,q0
-uint8x16_t vabdq_u8(uint8x16_t a, uint8x16_t b); // VABD.U8 q0,q0,q0
-uint16x8_t vabdq_u16(uint16x8_t a, uint16x8_t b); // VABD.U16 q0,q0,q0
-uint32x4_t vabdq_u32(uint32x4_t a, uint32x4_t b); // VABD.U32 q0,q0,q0
-float32x4_t vabdq_f32(float32x4_t a, float32x4_t b); // VABD.F32 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vabd_s8(int8x8_t a, int8x8_t b); // VABD.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vabd_s16(int16x4_t a, int16x4_t b); // VABD.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vabd_s32(int32x2_t a, int32x2_t b); // VABD.S32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vabd_u8(uint8x8_t a, uint8x8_t b); // VABD.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vabd_u16(uint16x4_t a, uint16x4_t b); // VABD.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vabd_u32(uint32x2_t a, uint32x2_t b); // VABD.U32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vabd_f32(float32x2_t a, float32x2_t b); // VABD.F32 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vabdq_s8(int8x16_t a, int8x16_t b); // VABD.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vabdq_s16(int16x8_t a, int16x8_t b); // VABD.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vabdq_s32(int32x4_t a, int32x4_t b); // VABD.S32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vabdq_u8(uint8x16_t a, uint8x16_t b); // VABD.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vabdq_u16(uint16x8_t a, uint16x8_t b); // VABD.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vabdq_u32(uint32x4_t a, uint32x4_t b); // VABD.U32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vabdq_f32(float32x4_t a, float32x4_t b); // VABD.F32 q0,q0,q0
 //Absolute difference - long
-int16x8_t vabdl_s8(int8x8_t a, int8x8_t b); // VABDL.S8 q0,d0,d0
-int32x4_t vabdl_s16(int16x4_t a, int16x4_t b); // VABDL.S16 q0,d0,d0
-int64x2_t vabdl_s32(int32x2_t a, int32x2_t b); // VABDL.S32 q0,d0,d0
-uint16x8_t vabdl_u8(uint8x8_t a, uint8x8_t b); // VABDL.U8 q0,d0,d0
-uint32x4_t vabdl_u16(uint16x4_t a, uint16x4_t b); // VABDL.U16 q0,d0,d0
-uint64x2_t vabdl_u32(uint32x2_t a, uint32x2_t b); // VABDL.U32 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vabdl_s8(int8x8_t a, int8x8_t b); // VABDL.S8 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vabdl_s16(int16x4_t a, int16x4_t b); // VABDL.S16 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vabdl_s32(int32x2_t a, int32x2_t b); // VABDL.S32 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vabdl_u8(uint8x8_t a, uint8x8_t b); // VABDL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vabdl_u16(uint16x4_t a, uint16x4_t b); // VABDL.U16 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vabdl_u32(uint32x2_t a, uint32x2_t b); // VABDL.U32 q0,d0,d0
 //Absolute difference and accumulate: Vr[i] = Va[i] + | Vb[i] - Vc[i] |
-int8x8_t vaba_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VABA.S8 d0,d0,d0
-int16x4_t vaba_s16(int16x4_t a, int16x4_t b, int16x4_t c); // VABA.S16 d0,d0,d0
-int32x2_t vaba_s32(int32x2_t a, int32x2_t b, int32x2_t c); // VABA.S32 d0,d0,d0
-uint8x8_t vaba_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VABA.U8 d0,d0,d0
-uint16x4_t vaba_u16(uint16x4_t a, uint16x4_t b, uint16x4_t c); // VABA.U16 d0,d0,d0
-uint32x2_t vaba_u32(uint32x2_t a, uint32x2_t b, uint32x2_t c); // VABA.U32 d0,d0,d0
-int8x16_t vabaq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VABA.S8 q0,q0,q0
-int16x8_t vabaq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VABA.S16 q0,q0,q0
-int32x4_t vabaq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VABA.S32 q0,q0,q0
-uint8x16_t vabaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VABA.U8 q0,q0,q0
-uint16x8_t vabaq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VABA.U16 q0,q0,q0
-uint32x4_t vabaq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VABA.U32 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vaba_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VABA.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vaba_s16(int16x4_t a, int16x4_t b, int16x4_t c); // VABA.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vaba_s32(int32x2_t a, int32x2_t b, int32x2_t c); // VABA.S32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vaba_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VABA.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vaba_u16(uint16x4_t a, uint16x4_t b, uint16x4_t c); // VABA.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vaba_u32(uint32x2_t a, uint32x2_t b, uint32x2_t c); // VABA.U32 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vabaq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VABA.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vabaq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VABA.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vabaq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VABA.S32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vabaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VABA.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vabaq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VABA.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vabaq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VABA.U32 q0,q0,q0
 //Absolute difference and accumulate - long
-int16x8_t vabal_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VABAL.S8 q0,d0,d0
-int32x4_t vabal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VABAL.S16 q0,d0,d0
-int64x2_t vabal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VABAL.S32 q0,d0,d0
-uint16x8_t vabal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VABAL.U8 q0,d0,d0
-uint32x4_t vabal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VABAL.U16 q0,d0,d0
-uint64x2_t vabal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VABAL.U32 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vabal_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VABAL.S8 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vabal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VABAL.S16 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vabal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VABAL.S32 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vabal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VABAL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vabal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VABAL.U16 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vabal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VABAL.U32 q0,d0,d0
 //Max/Min
 //vmax -> Vr[i] := (Va[i] >= Vb[i]) ? Va[i] : Vb[i]
-int8x8_t vmax_s8(int8x8_t a, int8x8_t b); // VMAX.S8 d0,d0,d0
-int16x4_t vmax_s16(int16x4_t a, int16x4_t b); // VMAX.S16 d0,d0,d0
-int32x2_t vmax_s32(int32x2_t a, int32x2_t b); // VMAX.S32 d0,d0,d0
-uint8x8_t vmax_u8(uint8x8_t a, uint8x8_t b); // VMAX.U8 d0,d0,d0
-uint16x4_t vmax_u16(uint16x4_t a, uint16x4_t b); // VMAX.U16 d0,d0,d0
-uint32x2_t vmax_u32(uint32x2_t a, uint32x2_t b); // VMAX.U32 d0,d0,d0
-float32x2_t vmax_f32(float32x2_t a, float32x2_t b); // VMAX.F32 d0,d0,d0
-int8x16_t vmaxq_s8(int8x16_t a, int8x16_t b); // VMAX.S8 q0,q0,q0
-int16x8_t vmaxq_s16(int16x8_t a, int16x8_t b); // VMAX.S16 q0,q0,q0
-int32x4_t vmaxq_s32(int32x4_t a, int32x4_t b); // VMAX.S32 q0,q0,q0
-uint8x16_t vmaxq_u8(uint8x16_t a, uint8x16_t b); // VMAX.U8 q0,q0,q0
-uint16x8_t vmaxq_u16(uint16x8_t a, uint16x8_t b); // VMAX.U16 q0,q0,q0
-uint32x4_t vmaxq_u32(uint32x4_t a, uint32x4_t b); // VMAX.U32 q0,q0,q0
-float32x4_t vmaxq_f32(float32x4_t a, float32x4_t b); // VMAX.F32 q0,q0,q0
-
-float64x2_t vmaxq_f64(float64x2_t a, float64x2_t b); // VMAX.F64 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vmax_s8(int8x8_t a, int8x8_t b); // VMAX.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vmax_s16(int16x4_t a, int16x4_t b); // VMAX.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vmax_s32(int32x2_t a, int32x2_t b); // VMAX.S32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vmax_u8(uint8x8_t a, uint8x8_t b); // VMAX.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vmax_u16(uint16x4_t a, uint16x4_t b); // VMAX.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vmax_u32(uint32x2_t a, uint32x2_t b); // VMAX.U32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vmax_f32(float32x2_t a, float32x2_t b); // VMAX.F32 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vmaxq_s8(int8x16_t a, int8x16_t b); // VMAX.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vmaxq_s16(int16x8_t a, int16x8_t b); // VMAX.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vmaxq_s32(int32x4_t a, int32x4_t b); // VMAX.S32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vmaxq_u8(uint8x16_t a, uint8x16_t b); // VMAX.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vmaxq_u16(uint16x8_t a, uint16x8_t b); // VMAX.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vmaxq_u32(uint32x4_t a, uint32x4_t b); // VMAX.U32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vmaxq_f32(float32x4_t a, float32x4_t b); // VMAX.F32 q0,q0,q0
+
+_NEON2SSESTORAGE float64x2_t vmaxq_f64(float64x2_t a, float64x2_t b); // VMAX.F64 q0,q0,q0
 
 //vmin -> Vr[i] := (Va[i] >= Vb[i]) ? Vb[i] : Va[i]
-int8x8_t vmin_s8(int8x8_t a, int8x8_t b); // VMIN.S8 d0,d0,d0
-int16x4_t vmin_s16(int16x4_t a, int16x4_t b); // VMIN.S16 d0,d0,d0
-int32x2_t vmin_s32(int32x2_t a, int32x2_t b); // VMIN.S32 d0,d0,d0
-uint8x8_t vmin_u8(uint8x8_t a, uint8x8_t b); // VMIN.U8 d0,d0,d0
-uint16x4_t vmin_u16(uint16x4_t a, uint16x4_t b); // VMIN.U16 d0,d0,d0
-uint32x2_t vmin_u32(uint32x2_t a, uint32x2_t b); // VMIN.U32 d0,d0,d0
-float32x2_t vmin_f32(float32x2_t a, float32x2_t b); // VMIN.F32 d0,d0,d0
-int8x16_t vminq_s8(int8x16_t a, int8x16_t b); // VMIN.S8 q0,q0,q0
-int16x8_t vminq_s16(int16x8_t a, int16x8_t b); // VMIN.S16 q0,q0,q0
-int32x4_t vminq_s32(int32x4_t a, int32x4_t b); // VMIN.S32 q0,q0,q0
-uint8x16_t vminq_u8(uint8x16_t a, uint8x16_t b); // VMIN.U8 q0,q0,q0
-uint16x8_t vminq_u16(uint16x8_t a, uint16x8_t b); // VMIN.U16 q0,q0,q0
-uint32x4_t vminq_u32(uint32x4_t a, uint32x4_t b); // VMIN.U32 q0,q0,q0
-float32x4_t vminq_f32(float32x4_t a, float32x4_t b); // VMIN.F32 q0,q0,q0
-
-float64x2_t vminq_f64(float64x2_t a, float64x2_t b); // VMIN.F64 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vmin_s8(int8x8_t a, int8x8_t b); // VMIN.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vmin_s16(int16x4_t a, int16x4_t b); // VMIN.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vmin_s32(int32x2_t a, int32x2_t b); // VMIN.S32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vmin_u8(uint8x8_t a, uint8x8_t b); // VMIN.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vmin_u16(uint16x4_t a, uint16x4_t b); // VMIN.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vmin_u32(uint32x2_t a, uint32x2_t b); // VMIN.U32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vmin_f32(float32x2_t a, float32x2_t b); // VMIN.F32 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vminq_s8(int8x16_t a, int8x16_t b); // VMIN.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vminq_s16(int16x8_t a, int16x8_t b); // VMIN.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vminq_s32(int32x4_t a, int32x4_t b); // VMIN.S32 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vminq_u8(uint8x16_t a, uint8x16_t b); // VMIN.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vminq_u16(uint16x8_t a, uint16x8_t b); // VMIN.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vminq_u32(uint32x4_t a, uint32x4_t b); // VMIN.U32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vminq_f32(float32x4_t a, float32x4_t b); // VMIN.F32 q0,q0,q0
+
+_NEON2SSESTORAGE float64x2_t vminq_f64(float64x2_t a, float64x2_t b); // VMIN.F64 q0,q0,q0
 
 //Pairwise addition
 //Pairwise add
-int8x8_t vpadd_s8(int8x8_t a, int8x8_t b); // VPADD.I8 d0,d0,d0
-int16x4_t vpadd_s16(int16x4_t a, int16x4_t b); // VPADD.I16 d0,d0,d0
-int32x2_t vpadd_s32(int32x2_t a, int32x2_t b); // VPADD.I32 d0,d0,d0
-uint8x8_t vpadd_u8(uint8x8_t a, uint8x8_t b); // VPADD.I8 d0,d0,d0
-uint16x4_t vpadd_u16(uint16x4_t a, uint16x4_t b); // VPADD.I16 d0,d0,d0
-uint32x2_t vpadd_u32(uint32x2_t a, uint32x2_t b); // VPADD.I32 d0,d0,d0
-float32x2_t vpadd_f32(float32x2_t a, float32x2_t b); // VPADD.F32 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vpadd_s8(int8x8_t a, int8x8_t b); // VPADD.I8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vpadd_s16(int16x4_t a, int16x4_t b); // VPADD.I16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vpadd_s32(int32x2_t a, int32x2_t b); // VPADD.I32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vpadd_u8(uint8x8_t a, uint8x8_t b); // VPADD.I8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vpadd_u16(uint16x4_t a, uint16x4_t b); // VPADD.I16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vpadd_u32(uint32x2_t a, uint32x2_t b); // VPADD.I32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vpadd_f32(float32x2_t a, float32x2_t b); // VPADD.F32 d0,d0,d0
 //Long pairwise add
-int16x4_t vpaddl_s8(int8x8_t a); // VPADDL.S8 d0,d0
-int32x2_t vpaddl_s16(int16x4_t a); // VPADDL.S16 d0,d0
-int64x1_t vpaddl_s32(int32x2_t a); // VPADDL.S32 d0,d0
-uint16x4_t vpaddl_u8(uint8x8_t a); // VPADDL.U8 d0,d0
-uint32x2_t vpaddl_u16(uint16x4_t a); // VPADDL.U16 d0,d0
-uint64x1_t vpaddl_u32(uint32x2_t a); // VPADDL.U32 d0,d0
-int16x8_t vpaddlq_s8(int8x16_t a); // VPADDL.S8 q0,q0
-int32x4_t vpaddlq_s16(int16x8_t a); // VPADDL.S16 q0,q0
-int64x2_t vpaddlq_s32(int32x4_t a); // VPADDL.S32 q0,q0
-uint16x8_t vpaddlq_u8(uint8x16_t a); // VPADDL.U8 q0,q0
-uint32x4_t vpaddlq_u16(uint16x8_t a); // VPADDL.U16 q0,q0
-uint64x2_t vpaddlq_u32(uint32x4_t a); // VPADDL.U32 q0,q0
+_NEON2SSESTORAGE int16x4_t vpaddl_s8(int8x8_t a); // VPADDL.S8 d0,d0
+_NEON2SSESTORAGE int32x2_t vpaddl_s16(int16x4_t a); // VPADDL.S16 d0,d0
+_NEON2SSESTORAGE int64x1_t vpaddl_s32(int32x2_t a); // VPADDL.S32 d0,d0
+_NEON2SSESTORAGE uint16x4_t vpaddl_u8(uint8x8_t a); // VPADDL.U8 d0,d0
+_NEON2SSESTORAGE uint32x2_t vpaddl_u16(uint16x4_t a); // VPADDL.U16 d0,d0
+_NEON2SSESTORAGE uint64x1_t vpaddl_u32(uint32x2_t a); // VPADDL.U32 d0,d0
+_NEON2SSESTORAGE int16x8_t vpaddlq_s8(int8x16_t a); // VPADDL.S8 q0,q0
+_NEON2SSESTORAGE int32x4_t vpaddlq_s16(int16x8_t a); // VPADDL.S16 q0,q0
+_NEON2SSESTORAGE int64x2_t vpaddlq_s32(int32x4_t a); // VPADDL.S32 q0,q0
+_NEON2SSESTORAGE uint16x8_t vpaddlq_u8(uint8x16_t a); // VPADDL.U8 q0,q0
+_NEON2SSESTORAGE uint32x4_t vpaddlq_u16(uint16x8_t a); // VPADDL.U16 q0,q0
+_NEON2SSESTORAGE uint64x2_t vpaddlq_u32(uint32x4_t a); // VPADDL.U32 q0,q0
 //Long pairwise add and accumulate
-int16x4_t vpadal_s8(int16x4_t a, int8x8_t b); // VPADAL.S8 d0,d0
-int32x2_t vpadal_s16(int32x2_t a, int16x4_t b); // VPADAL.S16 d0,d0
-int64x1_t vpadal_s32(int64x1_t a, int32x2_t b); // VPADAL.S32 d0,d0
-uint16x4_t vpadal_u8(uint16x4_t a, uint8x8_t b); // VPADAL.U8 d0,d0
-uint32x2_t vpadal_u16(uint32x2_t a, uint16x4_t b); // VPADAL.U16 d0,d0
-uint64x1_t vpadal_u32(uint64x1_t a, uint32x2_t b); // VPADAL.U32 d0,d0
-int16x8_t vpadalq_s8(int16x8_t a, int8x16_t b); // VPADAL.S8 q0,q0
-int32x4_t vpadalq_s16(int32x4_t a, int16x8_t b); // VPADAL.S16 q0,q0
-int64x2_t vpadalq_s32(int64x2_t a, int32x4_t b); // VPADAL.S32 q0,q0
-uint16x8_t vpadalq_u8(uint16x8_t a, uint8x16_t b); // VPADAL.U8 q0,q0
-uint32x4_t vpadalq_u16(uint32x4_t a, uint16x8_t b); // VPADAL.U16 q0,q0
-uint64x2_t vpadalq_u32(uint64x2_t a, uint32x4_t b); // VPADAL.U32 q0,q0
+_NEON2SSESTORAGE int16x4_t vpadal_s8(int16x4_t a, int8x8_t b); // VPADAL.S8 d0,d0
+_NEON2SSESTORAGE int32x2_t vpadal_s16(int32x2_t a, int16x4_t b); // VPADAL.S16 d0,d0
+_NEON2SSESTORAGE int64x1_t vpadal_s32(int64x1_t a, int32x2_t b); // VPADAL.S32 d0,d0
+_NEON2SSESTORAGE uint16x4_t vpadal_u8(uint16x4_t a, uint8x8_t b); // VPADAL.U8 d0,d0
+_NEON2SSESTORAGE uint32x2_t vpadal_u16(uint32x2_t a, uint16x4_t b); // VPADAL.U16 d0,d0
+_NEON2SSESTORAGE uint64x1_t vpadal_u32(uint64x1_t a, uint32x2_t b); // VPADAL.U32 d0,d0
+_NEON2SSESTORAGE int16x8_t vpadalq_s8(int16x8_t a, int8x16_t b); // VPADAL.S8 q0,q0
+_NEON2SSESTORAGE int32x4_t vpadalq_s16(int32x4_t a, int16x8_t b); // VPADAL.S16 q0,q0
+_NEON2SSESTORAGE int64x2_t vpadalq_s32(int64x2_t a, int32x4_t b); // VPADAL.S32 q0,q0
+_NEON2SSESTORAGE uint16x8_t vpadalq_u8(uint16x8_t a, uint8x16_t b); // VPADAL.U8 q0,q0
+_NEON2SSESTORAGE uint32x4_t vpadalq_u16(uint32x4_t a, uint16x8_t b); // VPADAL.U16 q0,q0
+_NEON2SSESTORAGE uint64x2_t vpadalq_u32(uint64x2_t a, uint32x4_t b); // VPADAL.U32 q0,q0
 //Folding maximum vpmax -> takes maximum of adjacent pairs
-int8x8_t vpmax_s8(int8x8_t a, int8x8_t b); // VPMAX.S8 d0,d0,d0
-int16x4_t vpmax_s16(int16x4_t a, int16x4_t b); // VPMAX.S16 d0,d0,d0
-int32x2_t vpmax_s32(int32x2_t a, int32x2_t b); // VPMAX.S32 d0,d0,d0
-uint8x8_t vpmax_u8(uint8x8_t a, uint8x8_t b); // VPMAX.U8 d0,d0,d0
-uint16x4_t vpmax_u16(uint16x4_t a, uint16x4_t b); // VPMAX.U16 d0,d0,d0
-uint32x2_t vpmax_u32(uint32x2_t a, uint32x2_t b); // VPMAX.U32 d0,d0,d0
-float32x2_t vpmax_f32(float32x2_t a, float32x2_t b); // VPMAX.F32 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vpmax_s8(int8x8_t a, int8x8_t b); // VPMAX.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vpmax_s16(int16x4_t a, int16x4_t b); // VPMAX.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vpmax_s32(int32x2_t a, int32x2_t b); // VPMAX.S32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vpmax_u8(uint8x8_t a, uint8x8_t b); // VPMAX.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vpmax_u16(uint16x4_t a, uint16x4_t b); // VPMAX.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vpmax_u32(uint32x2_t a, uint32x2_t b); // VPMAX.U32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vpmax_f32(float32x2_t a, float32x2_t b); // VPMAX.F32 d0,d0,d0
 //Folding minimum vpmin -> takes minimum of adjacent pairs
-int8x8_t vpmin_s8(int8x8_t a, int8x8_t b); // VPMIN.S8 d0,d0,d0
-int16x4_t vpmin_s16(int16x4_t a, int16x4_t b); // VPMIN.S16 d0,d0,d0
-int32x2_t vpmin_s32(int32x2_t a, int32x2_t b); // VPMIN.S32 d0,d0,d0
-uint8x8_t vpmin_u8(uint8x8_t a, uint8x8_t b); // VPMIN.U8 d0,d0,d0
-uint16x4_t vpmin_u16(uint16x4_t a, uint16x4_t b); // VPMIN.U16 d0,d0,d0
-uint32x2_t vpmin_u32(uint32x2_t a, uint32x2_t b); // VPMIN.U32 d0,d0,d0
-float32x2_t vpmin_f32(float32x2_t a, float32x2_t b); // VPMIN.F32 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vpmin_s8(int8x8_t a, int8x8_t b); // VPMIN.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vpmin_s16(int16x4_t a, int16x4_t b); // VPMIN.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vpmin_s32(int32x2_t a, int32x2_t b); // VPMIN.S32 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vpmin_u8(uint8x8_t a, uint8x8_t b); // VPMIN.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vpmin_u16(uint16x4_t a, uint16x4_t b); // VPMIN.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vpmin_u32(uint32x2_t a, uint32x2_t b); // VPMIN.U32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vpmin_f32(float32x2_t a, float32x2_t b); // VPMIN.F32 d0,d0,d0
 //Reciprocal/Sqrt
-float32x2_t vrecps_f32(float32x2_t a, float32x2_t b); // VRECPS.F32 d0, d0, d0
-float32x4_t vrecpsq_f32(float32x4_t a, float32x4_t b); // VRECPS.F32 q0, q0, q0
-float32x2_t vrsqrts_f32(float32x2_t a, float32x2_t b); // VRSQRTS.F32 d0, d0, d0
-float32x4_t vrsqrtsq_f32(float32x4_t a, float32x4_t b); // VRSQRTS.F32 q0, q0, q0
+_NEON2SSESTORAGE float32x2_t vrecps_f32(float32x2_t a, float32x2_t b); // VRECPS.F32 d0, d0, d0
+_NEON2SSESTORAGE float32x4_t vrecpsq_f32(float32x4_t a, float32x4_t b); // VRECPS.F32 q0, q0, q0
+_NEON2SSESTORAGE float32x2_t vrsqrts_f32(float32x2_t a, float32x2_t b); // VRSQRTS.F32 d0, d0, d0
+_NEON2SSESTORAGE float32x4_t vrsqrtsq_f32(float32x4_t a, float32x4_t b); // VRSQRTS.F32 q0, q0, q0
 //Shifts by signed variable
 //Vector shift left: Vr[i] := Va[i] << Vb[i] (negative values shift right)
-int8x8_t vshl_s8(int8x8_t a, int8x8_t b); // VSHL.S8 d0,d0,d0
-int16x4_t vshl_s16(int16x4_t a, int16x4_t b); // VSHL.S16 d0,d0,d0
-int32x2_t vshl_s32(int32x2_t a, int32x2_t b); // VSHL.S32 d0,d0,d0
-int64x1_t vshl_s64(int64x1_t a, int64x1_t b); // VSHL.S64 d0,d0,d0
-uint8x8_t vshl_u8(uint8x8_t a, int8x8_t b); // VSHL.U8 d0,d0,d0
-uint16x4_t vshl_u16(uint16x4_t a, int16x4_t b); // VSHL.U16 d0,d0,d0
-uint32x2_t vshl_u32(uint32x2_t a, int32x2_t b); // VSHL.U32 d0,d0,d0
-uint64x1_t vshl_u64(uint64x1_t a, int64x1_t b); // VSHL.U64 d0,d0,d0
-int8x16_t vshlq_s8(int8x16_t a, int8x16_t b); // VSHL.S8 q0,q0,q0
-int16x8_t vshlq_s16(int16x8_t a, int16x8_t b); // VSHL.S16 q0,q0,q0
-int32x4_t vshlq_s32(int32x4_t a, int32x4_t b); // VSHL.S32 q0,q0,q0
-int64x2_t vshlq_s64(int64x2_t a, int64x2_t b); // VSHL.S64 q0,q0,q0
-uint8x16_t vshlq_u8(uint8x16_t a, int8x16_t b); // VSHL.U8 q0,q0,q0
-uint16x8_t vshlq_u16(uint16x8_t a, int16x8_t b); // VSHL.U16 q0,q0,q0
-uint32x4_t vshlq_u32(uint32x4_t a, int32x4_t b); // VSHL.U32 q0,q0,q0
-uint64x2_t vshlq_u64(uint64x2_t a, int64x2_t b); // VSHL.U64 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vshl_s8(int8x8_t a, int8x8_t b); // VSHL.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vshl_s16(int16x4_t a, int16x4_t b); // VSHL.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vshl_s32(int32x2_t a, int32x2_t b); // VSHL.S32 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vshl_s64(int64x1_t a, int64x1_t b); // VSHL.S64 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vshl_u8(uint8x8_t a, int8x8_t b); // VSHL.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vshl_u16(uint16x4_t a, int16x4_t b); // VSHL.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vshl_u32(uint32x2_t a, int32x2_t b); // VSHL.U32 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vshl_u64(uint64x1_t a, int64x1_t b); // VSHL.U64 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vshlq_s8(int8x16_t a, int8x16_t b); // VSHL.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vshlq_s16(int16x8_t a, int16x8_t b); // VSHL.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vshlq_s32(int32x4_t a, int32x4_t b); // VSHL.S32 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vshlq_s64(int64x2_t a, int64x2_t b); // VSHL.S64 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vshlq_u8(uint8x16_t a, int8x16_t b); // VSHL.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vshlq_u16(uint16x8_t a, int16x8_t b); // VSHL.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vshlq_u32(uint32x4_t a, int32x4_t b); // VSHL.U32 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vshlq_u64(uint64x2_t a, int64x2_t b); // VSHL.U64 q0,q0,q0
 //Vector saturating shift left: (negative values shift right)
-int8x8_t vqshl_s8(int8x8_t a, int8x8_t b); // VQSHL.S8 d0,d0,d0
-int16x4_t vqshl_s16(int16x4_t a, int16x4_t b); // VQSHL.S16 d0,d0,d0
-int32x2_t vqshl_s32(int32x2_t a, int32x2_t b); // VQSHL.S32 d0,d0,d0
-int64x1_t vqshl_s64(int64x1_t a, int64x1_t b); // VQSHL.S64 d0,d0,d0
-uint8x8_t vqshl_u8(uint8x8_t a, int8x8_t b); // VQSHL.U8 d0,d0,d0
-uint16x4_t vqshl_u16(uint16x4_t a, int16x4_t b); // VQSHL.U16 d0,d0,d0
-uint32x2_t vqshl_u32(uint32x2_t a, int32x2_t b); // VQSHL.U32 d0,d0,d0
-uint64x1_t vqshl_u64(uint64x1_t a, int64x1_t b); // VQSHL.U64 d0,d0,d0
-int8x16_t vqshlq_s8(int8x16_t a, int8x16_t b); // VQSHL.S8 q0,q0,q0
-int16x8_t vqshlq_s16(int16x8_t a, int16x8_t b); // VQSHL.S16 q0,q0,q0
-int32x4_t vqshlq_s32(int32x4_t a, int32x4_t b); // VQSHL.S32 q0,q0,q0
-int64x2_t vqshlq_s64(int64x2_t a, int64x2_t b); // VQSHL.S64 q0,q0,q0
-uint8x16_t vqshlq_u8(uint8x16_t a, int8x16_t b); // VQSHL.U8 q0,q0,q0
-uint16x8_t vqshlq_u16(uint16x8_t a, int16x8_t b); // VQSHL.U16 q0,q0,q0
-uint32x4_t vqshlq_u32(uint32x4_t a, int32x4_t b); // VQSHL.U32 q0,q0,q0
-uint64x2_t vqshlq_u64(uint64x2_t a, int64x2_t b); // VQSHL.U64 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vqshl_s8(int8x8_t a, int8x8_t b); // VQSHL.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vqshl_s16(int16x4_t a, int16x4_t b); // VQSHL.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqshl_s32(int32x2_t a, int32x2_t b); // VQSHL.S32 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vqshl_s64(int64x1_t a, int64x1_t b); // VQSHL.S64 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vqshl_u8(uint8x8_t a, int8x8_t b); // VQSHL.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vqshl_u16(uint16x4_t a, int16x4_t b); // VQSHL.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vqshl_u32(uint32x2_t a, int32x2_t b); // VQSHL.U32 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vqshl_u64(uint64x1_t a, int64x1_t b); // VQSHL.U64 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vqshlq_s8(int8x16_t a, int8x16_t b); // VQSHL.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vqshlq_s16(int16x8_t a, int16x8_t b); // VQSHL.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vqshlq_s32(int32x4_t a, int32x4_t b); // VQSHL.S32 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vqshlq_s64(int64x2_t a, int64x2_t b); // VQSHL.S64 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vqshlq_u8(uint8x16_t a, int8x16_t b); // VQSHL.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vqshlq_u16(uint16x8_t a, int16x8_t b); // VQSHL.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vqshlq_u32(uint32x4_t a, int32x4_t b); // VQSHL.U32 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vqshlq_u64(uint64x2_t a, int64x2_t b); // VQSHL.U64 q0,q0,q0
 //Vector rounding shift left: (negative values shift right)
-int8x8_t vrshl_s8(int8x8_t a, int8x8_t b); // VRSHL.S8 d0,d0,d0
-int16x4_t vrshl_s16(int16x4_t a, int16x4_t b); // VRSHL.S16 d0,d0,d0
-int32x2_t vrshl_s32(int32x2_t a, int32x2_t b); // VRSHL.S32 d0,d0,d0
-int64x1_t vrshl_s64(int64x1_t a, int64x1_t b); // VRSHL.S64 d0,d0,d0
-uint8x8_t vrshl_u8(uint8x8_t a, int8x8_t b); // VRSHL.U8 d0,d0,d0
-uint16x4_t vrshl_u16(uint16x4_t a, int16x4_t b); // VRSHL.U16 d0,d0,d0
-uint32x2_t vrshl_u32(uint32x2_t a, int32x2_t b); // VRSHL.U32 d0,d0,d0
-uint64x1_t vrshl_u64(uint64x1_t a, int64x1_t b); // VRSHL.U64 d0,d0,d0
-int8x16_t vrshlq_s8(int8x16_t a, int8x16_t b); // VRSHL.S8 q0,q0,q0
-int16x8_t vrshlq_s16(int16x8_t a, int16x8_t b); // VRSHL.S16 q0,q0,q0
-int32x4_t vrshlq_s32(int32x4_t a, int32x4_t b); // VRSHL.S32 q0,q0,q0
-int64x2_t vrshlq_s64(int64x2_t a, int64x2_t b); // VRSHL.S64 q0,q0,q0
-uint8x16_t vrshlq_u8(uint8x16_t a, int8x16_t b); // VRSHL.U8 q0,q0,q0
-uint16x8_t vrshlq_u16(uint16x8_t a, int16x8_t b); // VRSHL.U16 q0,q0,q0
-uint32x4_t vrshlq_u32(uint32x4_t a, int32x4_t b); // VRSHL.U32 q0,q0,q0
-uint64x2_t vrshlq_u64(uint64x2_t a, int64x2_t b); // VRSHL.U64 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vrshl_s8(int8x8_t a, int8x8_t b); // VRSHL.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vrshl_s16(int16x4_t a, int16x4_t b); // VRSHL.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vrshl_s32(int32x2_t a, int32x2_t b); // VRSHL.S32 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vrshl_s64(int64x1_t a, int64x1_t b); // VRSHL.S64 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vrshl_u8(uint8x8_t a, int8x8_t b); // VRSHL.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vrshl_u16(uint16x4_t a, int16x4_t b); // VRSHL.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vrshl_u32(uint32x2_t a, int32x2_t b); // VRSHL.U32 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vrshl_u64(uint64x1_t a, int64x1_t b); // VRSHL.U64 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vrshlq_s8(int8x16_t a, int8x16_t b); // VRSHL.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vrshlq_s16(int16x8_t a, int16x8_t b); // VRSHL.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vrshlq_s32(int32x4_t a, int32x4_t b); // VRSHL.S32 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vrshlq_s64(int64x2_t a, int64x2_t b); // VRSHL.S64 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vrshlq_u8(uint8x16_t a, int8x16_t b); // VRSHL.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vrshlq_u16(uint16x8_t a, int16x8_t b); // VRSHL.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vrshlq_u32(uint32x4_t a, int32x4_t b); // VRSHL.U32 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vrshlq_u64(uint64x2_t a, int64x2_t b); // VRSHL.U64 q0,q0,q0
 //Vector saturating rounding shift left: (negative values shift right)
-int8x8_t vqrshl_s8(int8x8_t a, int8x8_t b); // VQRSHL.S8 d0,d0,d0
-int16x4_t vqrshl_s16(int16x4_t a, int16x4_t b); // VQRSHL.S16 d0,d0,d0
-int32x2_t vqrshl_s32(int32x2_t a, int32x2_t b); // VQRSHL.S32 d0,d0,d0
-int64x1_t vqrshl_s64(int64x1_t a, int64x1_t b); // VQRSHL.S64 d0,d0,d0
-uint8x8_t vqrshl_u8(uint8x8_t a, int8x8_t b); // VQRSHL.U8 d0,d0,d0
-uint16x4_t vqrshl_u16(uint16x4_t a, int16x4_t b); // VQRSHL.U16 d0,d0,d0
-uint32x2_t vqrshl_u32(uint32x2_t a, int32x2_t b); // VQRSHL.U32 d0,d0,d0
-uint64x1_t vqrshl_u64(uint64x1_t a, int64x1_t b); // VQRSHL.U64 d0,d0,d0
-int8x16_t vqrshlq_s8(int8x16_t a, int8x16_t b); // VQRSHL.S8 q0,q0,q0
-int16x8_t vqrshlq_s16(int16x8_t a, int16x8_t b); // VQRSHL.S16 q0,q0,q0
-int32x4_t vqrshlq_s32(int32x4_t a, int32x4_t b); // VQRSHL.S32 q0,q0,q0
-int64x2_t vqrshlq_s64(int64x2_t a, int64x2_t b); // VQRSHL.S64 q0,q0,q0
-uint8x16_t vqrshlq_u8(uint8x16_t a, int8x16_t b); // VQRSHL.U8 q0,q0,q0
-uint16x8_t vqrshlq_u16(uint16x8_t a, int16x8_t b); // VQRSHL.U16 q0,q0,q0
-uint32x4_t vqrshlq_u32(uint32x4_t a, int32x4_t b); // VQRSHL.U32 q0,q0,q0
-uint64x2_t vqrshlq_u64(uint64x2_t a, int64x2_t b); // VQRSHL.U64 q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vqrshl_s8(int8x8_t a, int8x8_t b); // VQRSHL.S8 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vqrshl_s16(int16x4_t a, int16x4_t b); // VQRSHL.S16 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqrshl_s32(int32x2_t a, int32x2_t b); // VQRSHL.S32 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vqrshl_s64(int64x1_t a, int64x1_t b); // VQRSHL.S64 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vqrshl_u8(uint8x8_t a, int8x8_t b); // VQRSHL.U8 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vqrshl_u16(uint16x4_t a, int16x4_t b); // VQRSHL.U16 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vqrshl_u32(uint32x2_t a, int32x2_t b); // VQRSHL.U32 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vqrshl_u64(uint64x1_t a, int64x1_t b); // VQRSHL.U64 d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vqrshlq_s8(int8x16_t a, int8x16_t b); // VQRSHL.S8 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vqrshlq_s16(int16x8_t a, int16x8_t b); // VQRSHL.S16 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vqrshlq_s32(int32x4_t a, int32x4_t b); // VQRSHL.S32 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vqrshlq_s64(int64x2_t a, int64x2_t b); // VQRSHL.S64 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vqrshlq_u8(uint8x16_t a, int8x16_t b); // VQRSHL.U8 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vqrshlq_u16(uint16x8_t a, int16x8_t b); // VQRSHL.U16 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vqrshlq_u32(uint32x4_t a, int32x4_t b); // VQRSHL.U32 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vqrshlq_u64(uint64x2_t a, int64x2_t b); // VQRSHL.U64 q0,q0,q0
 //Shifts by a constant
 //Vector shift right by constant
-int8x8_t vshr_n_s8(int8x8_t a, __constrange(1,8) int b); // VSHR.S8 d0,d0,#8
-int16x4_t vshr_n_s16(int16x4_t a, __constrange(1,16) int b); // VSHR.S16 d0,d0,#16
-int32x2_t vshr_n_s32(int32x2_t a, __constrange(1,32) int b); // VSHR.S32 d0,d0,#32
-int64x1_t vshr_n_s64(int64x1_t a, __constrange(1,64) int b); // VSHR.S64 d0,d0,#64
-uint8x8_t vshr_n_u8(uint8x8_t a, __constrange(1,8) int b); // VSHR.U8 d0,d0,#8
-uint16x4_t vshr_n_u16(uint16x4_t a, __constrange(1,16) int b); // VSHR.U16 d0,d0,#16
-uint32x2_t vshr_n_u32(uint32x2_t a, __constrange(1,32) int b); // VSHR.U32 d0,d0,#32
-uint64x1_t vshr_n_u64(uint64x1_t a, __constrange(1,64) int b); // VSHR.U64 d0,d0,#64
-int8x16_t vshrq_n_s8(int8x16_t a, __constrange(1,8) int b); // VSHR.S8 q0,q0,#8
-int16x8_t vshrq_n_s16(int16x8_t a, __constrange(1,16) int b); // VSHR.S16 q0,q0,#16
-int32x4_t vshrq_n_s32(int32x4_t a, __constrange(1,32) int b); // VSHR.S32 q0,q0,#32
-int64x2_t vshrq_n_s64(int64x2_t a, __constrange(1,64) int b); // VSHR.S64 q0,q0,#64
-uint8x16_t vshrq_n_u8(uint8x16_t a, __constrange(1,8) int b); // VSHR.U8 q0,q0,#8
-uint16x8_t vshrq_n_u16(uint16x8_t a, __constrange(1,16) int b); // VSHR.U16 q0,q0,#16
-uint32x4_t vshrq_n_u32(uint32x4_t a, __constrange(1,32) int b); // VSHR.U32 q0,q0,#32
-uint64x2_t vshrq_n_u64(uint64x2_t a, __constrange(1,64) int b); // VSHR.U64 q0,q0,#64
+_NEON2SSESTORAGE int8x8_t vshr_n_s8(int8x8_t a, __constrange(1,8) int b); // VSHR.S8 d0,d0,#8
+_NEON2SSESTORAGE int16x4_t vshr_n_s16(int16x4_t a, __constrange(1,16) int b); // VSHR.S16 d0,d0,#16
+_NEON2SSESTORAGE int32x2_t vshr_n_s32(int32x2_t a, __constrange(1,32) int b); // VSHR.S32 d0,d0,#32
+_NEON2SSESTORAGE int64x1_t vshr_n_s64(int64x1_t a, __constrange(1,64) int b); // VSHR.S64 d0,d0,#64
+_NEON2SSESTORAGE uint8x8_t vshr_n_u8(uint8x8_t a, __constrange(1,8) int b); // VSHR.U8 d0,d0,#8
+_NEON2SSESTORAGE uint16x4_t vshr_n_u16(uint16x4_t a, __constrange(1,16) int b); // VSHR.U16 d0,d0,#16
+_NEON2SSESTORAGE uint32x2_t vshr_n_u32(uint32x2_t a, __constrange(1,32) int b); // VSHR.U32 d0,d0,#32
+_NEON2SSESTORAGE uint64x1_t vshr_n_u64(uint64x1_t a, __constrange(1,64) int b); // VSHR.U64 d0,d0,#64
+_NEON2SSESTORAGE int8x16_t vshrq_n_s8(int8x16_t a, __constrange(1,8) int b); // VSHR.S8 q0,q0,#8
+_NEON2SSESTORAGE int16x8_t vshrq_n_s16(int16x8_t a, __constrange(1,16) int b); // VSHR.S16 q0,q0,#16
+_NEON2SSESTORAGE int32x4_t vshrq_n_s32(int32x4_t a, __constrange(1,32) int b); // VSHR.S32 q0,q0,#32
+_NEON2SSESTORAGE int64x2_t vshrq_n_s64(int64x2_t a, __constrange(1,64) int b); // VSHR.S64 q0,q0,#64
+_NEON2SSESTORAGE uint8x16_t vshrq_n_u8(uint8x16_t a, __constrange(1,8) int b); // VSHR.U8 q0,q0,#8
+_NEON2SSESTORAGE uint16x8_t vshrq_n_u16(uint16x8_t a, __constrange(1,16) int b); // VSHR.U16 q0,q0,#16
+_NEON2SSESTORAGE uint32x4_t vshrq_n_u32(uint32x4_t a, __constrange(1,32) int b); // VSHR.U32 q0,q0,#32
+_NEON2SSESTORAGE uint64x2_t vshrq_n_u64(uint64x2_t a, __constrange(1,64) int b); // VSHR.U64 q0,q0,#64
 //Vector shift left by constant
-int8x8_t vshl_n_s8(int8x8_t a, __constrange(0,7) int b); // VSHL.I8 d0,d0,#0
-int16x4_t vshl_n_s16(int16x4_t a, __constrange(0,15) int b); // VSHL.I16 d0,d0,#0
-int32x2_t vshl_n_s32(int32x2_t a, __constrange(0,31) int b); // VSHL.I32 d0,d0,#0
-int64x1_t vshl_n_s64(int64x1_t a, __constrange(0,63) int b); // VSHL.I64 d0,d0,#0
-uint8x8_t vshl_n_u8(uint8x8_t a, __constrange(0,7) int b); // VSHL.I8 d0,d0,#0
-uint16x4_t vshl_n_u16(uint16x4_t a, __constrange(0,15) int b); // VSHL.I16 d0,d0,#0
-uint32x2_t vshl_n_u32(uint32x2_t a, __constrange(0,31) int b); // VSHL.I32 d0,d0,#0
-uint64x1_t vshl_n_u64(uint64x1_t a, __constrange(0,63) int b); // VSHL.I64 d0,d0,#0
-int8x16_t vshlq_n_s8(int8x16_t a, __constrange(0,7) int b); // VSHL.I8 q0,q0,#0
-int16x8_t vshlq_n_s16(int16x8_t a, __constrange(0,15) int b); // VSHL.I16 q0,q0,#0
-int32x4_t vshlq_n_s32(int32x4_t a, __constrange(0,31) int b); // VSHL.I32 q0,q0,#0
-int64x2_t vshlq_n_s64(int64x2_t a, __constrange(0,63) int b); // VSHL.I64 q0,q0,#0
-uint8x16_t vshlq_n_u8(uint8x16_t a, __constrange(0,7) int b); // VSHL.I8 q0,q0,#0
-uint16x8_t vshlq_n_u16(uint16x8_t a, __constrange(0,15) int b); // VSHL.I16 q0,q0,#0
-uint32x4_t vshlq_n_u32(uint32x4_t a, __constrange(0,31) int b); // VSHL.I32 q0,q0,#0
-uint64x2_t vshlq_n_u64(uint64x2_t a, __constrange(0,63) int b); // VSHL.I64 q0,q0,#0
+_NEON2SSESTORAGE int8x8_t vshl_n_s8(int8x8_t a, __constrange(0,7) int b); // VSHL.I8 d0,d0,#0
+_NEON2SSESTORAGE int16x4_t vshl_n_s16(int16x4_t a, __constrange(0,15) int b); // VSHL.I16 d0,d0,#0
+_NEON2SSESTORAGE int32x2_t vshl_n_s32(int32x2_t a, __constrange(0,31) int b); // VSHL.I32 d0,d0,#0
+_NEON2SSESTORAGE int64x1_t vshl_n_s64(int64x1_t a, __constrange(0,63) int b); // VSHL.I64 d0,d0,#0
+_NEON2SSESTORAGE uint8x8_t vshl_n_u8(uint8x8_t a, __constrange(0,7) int b); // VSHL.I8 d0,d0,#0
+_NEON2SSESTORAGE uint16x4_t vshl_n_u16(uint16x4_t a, __constrange(0,15) int b); // VSHL.I16 d0,d0,#0
+_NEON2SSESTORAGE uint32x2_t vshl_n_u32(uint32x2_t a, __constrange(0,31) int b); // VSHL.I32 d0,d0,#0
+_NEON2SSESTORAGE uint64x1_t vshl_n_u64(uint64x1_t a, __constrange(0,63) int b); // VSHL.I64 d0,d0,#0
+_NEON2SSESTORAGE int8x16_t vshlq_n_s8(int8x16_t a, __constrange(0,7) int b); // VSHL.I8 q0,q0,#0
+_NEON2SSESTORAGE int16x8_t vshlq_n_s16(int16x8_t a, __constrange(0,15) int b); // VSHL.I16 q0,q0,#0
+_NEON2SSESTORAGE int32x4_t vshlq_n_s32(int32x4_t a, __constrange(0,31) int b); // VSHL.I32 q0,q0,#0
+_NEON2SSESTORAGE int64x2_t vshlq_n_s64(int64x2_t a, __constrange(0,63) int b); // VSHL.I64 q0,q0,#0
+_NEON2SSESTORAGE uint8x16_t vshlq_n_u8(uint8x16_t a, __constrange(0,7) int b); // VSHL.I8 q0,q0,#0
+_NEON2SSESTORAGE uint16x8_t vshlq_n_u16(uint16x8_t a, __constrange(0,15) int b); // VSHL.I16 q0,q0,#0
+_NEON2SSESTORAGE uint32x4_t vshlq_n_u32(uint32x4_t a, __constrange(0,31) int b); // VSHL.I32 q0,q0,#0
+_NEON2SSESTORAGE uint64x2_t vshlq_n_u64(uint64x2_t a, __constrange(0,63) int b); // VSHL.I64 q0,q0,#0
 //Vector rounding shift right by constant
-int8x8_t vrshr_n_s8(int8x8_t a, __constrange(1,8) int b); // VRSHR.S8 d0,d0,#8
-int16x4_t vrshr_n_s16(int16x4_t a, __constrange(1,16) int b); // VRSHR.S16 d0,d0,#16
-int32x2_t vrshr_n_s32(int32x2_t a, __constrange(1,32) int b); // VRSHR.S32 d0,d0,#32
-int64x1_t vrshr_n_s64(int64x1_t a, __constrange(1,64) int b); // VRSHR.S64 d0,d0,#64
-uint8x8_t vrshr_n_u8(uint8x8_t a, __constrange(1,8) int b); // VRSHR.U8 d0,d0,#8
-uint16x4_t vrshr_n_u16(uint16x4_t a, __constrange(1,16) int b); // VRSHR.U16 d0,d0,#16
-uint32x2_t vrshr_n_u32(uint32x2_t a, __constrange(1,32) int b); // VRSHR.U32 d0,d0,#32
-uint64x1_t vrshr_n_u64(uint64x1_t a, __constrange(1,64) int b); // VRSHR.U64 d0,d0,#64
-int8x16_t vrshrq_n_s8(int8x16_t a, __constrange(1,8) int b); // VRSHR.S8 q0,q0,#8
-int16x8_t vrshrq_n_s16(int16x8_t a, __constrange(1,16) int b); // VRSHR.S16 q0,q0,#16
-int32x4_t vrshrq_n_s32(int32x4_t a, __constrange(1,32) int b); // VRSHR.S32 q0,q0,#32
-int64x2_t vrshrq_n_s64(int64x2_t a, __constrange(1,64) int b); // VRSHR.S64 q0,q0,#64
-uint8x16_t vrshrq_n_u8(uint8x16_t a, __constrange(1,8) int b); // VRSHR.U8 q0,q0,#8
-uint16x8_t vrshrq_n_u16(uint16x8_t a, __constrange(1,16) int b); // VRSHR.U16 q0,q0,#16
-uint32x4_t vrshrq_n_u32(uint32x4_t a, __constrange(1,32) int b); // VRSHR.U32 q0,q0,#32
-uint64x2_t vrshrq_n_u64(uint64x2_t a, __constrange(1,64) int b); // VRSHR.U64 q0,q0,#64
+_NEON2SSESTORAGE int8x8_t vrshr_n_s8(int8x8_t a, __constrange(1,8) int b); // VRSHR.S8 d0,d0,#8
+_NEON2SSESTORAGE int16x4_t vrshr_n_s16(int16x4_t a, __constrange(1,16) int b); // VRSHR.S16 d0,d0,#16
+_NEON2SSESTORAGE int32x2_t vrshr_n_s32(int32x2_t a, __constrange(1,32) int b); // VRSHR.S32 d0,d0,#32
+_NEON2SSESTORAGE int64x1_t vrshr_n_s64(int64x1_t a, __constrange(1,64) int b); // VRSHR.S64 d0,d0,#64
+_NEON2SSESTORAGE uint8x8_t vrshr_n_u8(uint8x8_t a, __constrange(1,8) int b); // VRSHR.U8 d0,d0,#8
+_NEON2SSESTORAGE uint16x4_t vrshr_n_u16(uint16x4_t a, __constrange(1,16) int b); // VRSHR.U16 d0,d0,#16
+_NEON2SSESTORAGE uint32x2_t vrshr_n_u32(uint32x2_t a, __constrange(1,32) int b); // VRSHR.U32 d0,d0,#32
+_NEON2SSESTORAGE uint64x1_t vrshr_n_u64(uint64x1_t a, __constrange(1,64) int b); // VRSHR.U64 d0,d0,#64
+_NEON2SSESTORAGE int8x16_t vrshrq_n_s8(int8x16_t a, __constrange(1,8) int b); // VRSHR.S8 q0,q0,#8
+_NEON2SSESTORAGE int16x8_t vrshrq_n_s16(int16x8_t a, __constrange(1,16) int b); // VRSHR.S16 q0,q0,#16
+_NEON2SSESTORAGE int32x4_t vrshrq_n_s32(int32x4_t a, __constrange(1,32) int b); // VRSHR.S32 q0,q0,#32
+_NEON2SSESTORAGE int64x2_t vrshrq_n_s64(int64x2_t a, __constrange(1,64) int b); // VRSHR.S64 q0,q0,#64
+_NEON2SSESTORAGE uint8x16_t vrshrq_n_u8(uint8x16_t a, __constrange(1,8) int b); // VRSHR.U8 q0,q0,#8
+_NEON2SSESTORAGE uint16x8_t vrshrq_n_u16(uint16x8_t a, __constrange(1,16) int b); // VRSHR.U16 q0,q0,#16
+_NEON2SSESTORAGE uint32x4_t vrshrq_n_u32(uint32x4_t a, __constrange(1,32) int b); // VRSHR.U32 q0,q0,#32
+_NEON2SSESTORAGE uint64x2_t vrshrq_n_u64(uint64x2_t a, __constrange(1,64) int b); // VRSHR.U64 q0,q0,#64
 //Vector shift right by constant and accumulate
-int8x8_t vsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c); // VSRA.S8 d0,d0,#8
-int16x4_t vsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c); // VSRA.S16 d0,d0,#16
-int32x2_t vsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c); // VSRA.S32 d0,d0,#32
-int64x1_t vsra_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VSRA.S64 d0,d0,#64
-uint8x8_t vsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c); // VSRA.U8 d0,d0,#8
-uint16x4_t vsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c); // VSRA.U16 d0,d0,#16
-uint32x2_t vsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c); // VSRA.U32 d0,d0,#32
-uint64x1_t vsra_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VSRA.U64 d0,d0,#64
-int8x16_t vsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VSRA.S8 q0,q0,#8
-int16x8_t vsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VSRA.S16 q0,q0,#16
-int32x4_t vsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VSRA.S32 q0,q0,#32
-int64x2_t vsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VSRA.S64 q0,q0,#64
-uint8x16_t vsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VSRA.U8 q0,q0,#8
-uint16x8_t vsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VSRA.U16 q0,q0,#16
-uint32x4_t vsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VSRA.U32 q0,q0,#32
-uint64x2_t vsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VSRA.U64 q0,q0,#64
+_NEON2SSESTORAGE int8x8_t vsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c); // VSRA.S8 d0,d0,#8
+_NEON2SSESTORAGE int16x4_t vsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c); // VSRA.S16 d0,d0,#16
+_NEON2SSESTORAGE int32x2_t vsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c); // VSRA.S32 d0,d0,#32
+_NEON2SSESTORAGE int64x1_t vsra_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VSRA.S64 d0,d0,#64
+_NEON2SSESTORAGE uint8x8_t vsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c); // VSRA.U8 d0,d0,#8
+_NEON2SSESTORAGE uint16x4_t vsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c); // VSRA.U16 d0,d0,#16
+_NEON2SSESTORAGE uint32x2_t vsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c); // VSRA.U32 d0,d0,#32
+_NEON2SSESTORAGE uint64x1_t vsra_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VSRA.U64 d0,d0,#64
+_NEON2SSESTORAGE int8x16_t vsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VSRA.S8 q0,q0,#8
+_NEON2SSESTORAGE int16x8_t vsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VSRA.S16 q0,q0,#16
+_NEON2SSESTORAGE int32x4_t vsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VSRA.S32 q0,q0,#32
+_NEON2SSESTORAGE int64x2_t vsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VSRA.S64 q0,q0,#64
+_NEON2SSESTORAGE uint8x16_t vsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VSRA.U8 q0,q0,#8
+_NEON2SSESTORAGE uint16x8_t vsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VSRA.U16 q0,q0,#16
+_NEON2SSESTORAGE uint32x4_t vsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VSRA.U32 q0,q0,#32
+_NEON2SSESTORAGE uint64x2_t vsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VSRA.U64 q0,q0,#64
 //Vector rounding shift right by constant and accumulate
-int8x8_t vrsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c); // VRSRA.S8 d0,d0,#8
-int16x4_t vrsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c); // VRSRA.S16 d0,d0,#16
-int32x2_t vrsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c); // VRSRA.S32 d0,d0,#32
-int64x1_t vrsra_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VRSRA.S64 d0,d0,#64
-uint8x8_t vrsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c); // VRSRA.U8 d0,d0,#8
-uint16x4_t vrsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c); // VRSRA.U16 d0,d0,#16
-uint32x2_t vrsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c); // VRSRA.U32 d0,d0,#32
-uint64x1_t vrsra_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VRSRA.U64 d0,d0,#64
-int8x16_t vrsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VRSRA.S8 q0,q0,#8
-int16x8_t vrsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VRSRA.S16 q0,q0,#16
-int32x4_t vrsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VRSRA.S32 q0,q0,#32
-int64x2_t vrsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VRSRA.S64 q0,q0,#64
-uint8x16_t vrsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VRSRA.U8 q0,q0,#8
-uint16x8_t vrsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VRSRA.U16 q0,q0,#16
-uint32x4_t vrsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VRSRA.U32 q0,q0,#32
-uint64x2_t vrsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VRSRA.U64 q0,q0,#64
+_NEON2SSESTORAGE int8x8_t vrsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c); // VRSRA.S8 d0,d0,#8
+_NEON2SSESTORAGE int16x4_t vrsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c); // VRSRA.S16 d0,d0,#16
+_NEON2SSESTORAGE int32x2_t vrsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c); // VRSRA.S32 d0,d0,#32
+_NEON2SSESTORAGE int64x1_t vrsra_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VRSRA.S64 d0,d0,#64
+_NEON2SSESTORAGE uint8x8_t vrsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c); // VRSRA.U8 d0,d0,#8
+_NEON2SSESTORAGE uint16x4_t vrsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c); // VRSRA.U16 d0,d0,#16
+_NEON2SSESTORAGE uint32x2_t vrsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c); // VRSRA.U32 d0,d0,#32
+_NEON2SSESTORAGE uint64x1_t vrsra_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VRSRA.U64 d0,d0,#64
+_NEON2SSESTORAGE int8x16_t vrsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VRSRA.S8 q0,q0,#8
+_NEON2SSESTORAGE int16x8_t vrsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VRSRA.S16 q0,q0,#16
+_NEON2SSESTORAGE int32x4_t vrsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VRSRA.S32 q0,q0,#32
+_NEON2SSESTORAGE int64x2_t vrsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VRSRA.S64 q0,q0,#64
+_NEON2SSESTORAGE uint8x16_t vrsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VRSRA.U8 q0,q0,#8
+_NEON2SSESTORAGE uint16x8_t vrsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VRSRA.U16 q0,q0,#16
+_NEON2SSESTORAGE uint32x4_t vrsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VRSRA.U32 q0,q0,#32
+_NEON2SSESTORAGE uint64x2_t vrsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VRSRA.U64 q0,q0,#64
 //Vector saturating shift left by constant
-int8x8_t vqshl_n_s8(int8x8_t a, __constrange(0,7) int b); // VQSHL.S8 d0,d0,#0
-int16x4_t vqshl_n_s16(int16x4_t a, __constrange(0,15) int b); // VQSHL.S16 d0,d0,#0
-int32x2_t vqshl_n_s32(int32x2_t a, __constrange(0,31) int b); // VQSHL.S32 d0,d0,#0
-int64x1_t vqshl_n_s64(int64x1_t a, __constrange(0,63) int b); // VQSHL.S64 d0,d0,#0
-uint8x8_t vqshl_n_u8(uint8x8_t a, __constrange(0,7) int b); // VQSHL.U8 d0,d0,#0
-uint16x4_t vqshl_n_u16(uint16x4_t a, __constrange(0,15) int b); // VQSHL.U16 d0,d0,#0
-uint32x2_t vqshl_n_u32(uint32x2_t a, __constrange(0,31) int b); // VQSHL.U32 d0,d0,#0
-uint64x1_t vqshl_n_u64(uint64x1_t a, __constrange(0,63) int b); // VQSHL.U64 d0,d0,#0
-int8x16_t vqshlq_n_s8(int8x16_t a, __constrange(0,7) int b); // VQSHL.S8 q0,q0,#0
-int16x8_t vqshlq_n_s16(int16x8_t a, __constrange(0,15) int b); // VQSHL.S16 q0,q0,#0
-int32x4_t vqshlq_n_s32(int32x4_t a, __constrange(0,31) int b); // VQSHL.S32 q0,q0,#0
-int64x2_t vqshlq_n_s64(int64x2_t a, __constrange(0,63) int b); // VQSHL.S64 q0,q0,#0
-uint8x16_t vqshlq_n_u8(uint8x16_t a, __constrange(0,7) int b); // VQSHL.U8 q0,q0,#0
-uint16x8_t vqshlq_n_u16(uint16x8_t a, __constrange(0,15) int b); // VQSHL.U16 q0,q0,#0
-uint32x4_t vqshlq_n_u32(uint32x4_t a, __constrange(0,31) int b); // VQSHL.U32 q0,q0,#0
-uint64x2_t vqshlq_n_u64(uint64x2_t a, __constrange(0,63) int b); // VQSHL.U64 q0,q0,#0
+_NEON2SSESTORAGE int8x8_t vqshl_n_s8(int8x8_t a, __constrange(0,7) int b); // VQSHL.S8 d0,d0,#0
+_NEON2SSESTORAGE int16x4_t vqshl_n_s16(int16x4_t a, __constrange(0,15) int b); // VQSHL.S16 d0,d0,#0
+_NEON2SSESTORAGE int32x2_t vqshl_n_s32(int32x2_t a, __constrange(0,31) int b); // VQSHL.S32 d0,d0,#0
+_NEON2SSESTORAGE int64x1_t vqshl_n_s64(int64x1_t a, __constrange(0,63) int b); // VQSHL.S64 d0,d0,#0
+_NEON2SSESTORAGE uint8x8_t vqshl_n_u8(uint8x8_t a, __constrange(0,7) int b); // VQSHL.U8 d0,d0,#0
+_NEON2SSESTORAGE uint16x4_t vqshl_n_u16(uint16x4_t a, __constrange(0,15) int b); // VQSHL.U16 d0,d0,#0
+_NEON2SSESTORAGE uint32x2_t vqshl_n_u32(uint32x2_t a, __constrange(0,31) int b); // VQSHL.U32 d0,d0,#0
+_NEON2SSESTORAGE uint64x1_t vqshl_n_u64(uint64x1_t a, __constrange(0,63) int b); // VQSHL.U64 d0,d0,#0
+_NEON2SSESTORAGE int8x16_t vqshlq_n_s8(int8x16_t a, __constrange(0,7) int b); // VQSHL.S8 q0,q0,#0
+_NEON2SSESTORAGE int16x8_t vqshlq_n_s16(int16x8_t a, __constrange(0,15) int b); // VQSHL.S16 q0,q0,#0
+_NEON2SSESTORAGE int32x4_t vqshlq_n_s32(int32x4_t a, __constrange(0,31) int b); // VQSHL.S32 q0,q0,#0
+_NEON2SSESTORAGE int64x2_t vqshlq_n_s64(int64x2_t a, __constrange(0,63) int b); // VQSHL.S64 q0,q0,#0
+_NEON2SSESTORAGE uint8x16_t vqshlq_n_u8(uint8x16_t a, __constrange(0,7) int b); // VQSHL.U8 q0,q0,#0
+_NEON2SSESTORAGE uint16x8_t vqshlq_n_u16(uint16x8_t a, __constrange(0,15) int b); // VQSHL.U16 q0,q0,#0
+_NEON2SSESTORAGE uint32x4_t vqshlq_n_u32(uint32x4_t a, __constrange(0,31) int b); // VQSHL.U32 q0,q0,#0
+_NEON2SSESTORAGE uint64x2_t vqshlq_n_u64(uint64x2_t a, __constrange(0,63) int b); // VQSHL.U64 q0,q0,#0
 //Vector signed->unsigned saturating shift left by constant
-uint8x8_t vqshlu_n_s8(int8x8_t a, __constrange(0,7) int b); // VQSHLU.S8 d0,d0,#0
-uint16x4_t vqshlu_n_s16(int16x4_t a, __constrange(0,15) int b); // VQSHLU.S16 d0,d0,#0
-uint32x2_t vqshlu_n_s32(int32x2_t a, __constrange(0,31) int b); // VQSHLU.S32 d0,d0,#0
-uint64x1_t vqshlu_n_s64(int64x1_t a, __constrange(0,63) int b); // VQSHLU.S64 d0,d0,#0
-uint8x16_t vqshluq_n_s8(int8x16_t a, __constrange(0,7) int b); // VQSHLU.S8 q0,q0,#0
-uint16x8_t vqshluq_n_s16(int16x8_t a, __constrange(0,15) int b); // VQSHLU.S16 q0,q0,#0
-uint32x4_t vqshluq_n_s32(int32x4_t a, __constrange(0,31) int b); // VQSHLU.S32 q0,q0,#0
-uint64x2_t vqshluq_n_s64(int64x2_t a, __constrange(0,63) int b); // VQSHLU.S64 q0,q0,#0
+_NEON2SSESTORAGE uint8x8_t vqshlu_n_s8(int8x8_t a, __constrange(0,7) int b); // VQSHLU.S8 d0,d0,#0
+_NEON2SSESTORAGE uint16x4_t vqshlu_n_s16(int16x4_t a, __constrange(0,15) int b); // VQSHLU.S16 d0,d0,#0
+_NEON2SSESTORAGE uint32x2_t vqshlu_n_s32(int32x2_t a, __constrange(0,31) int b); // VQSHLU.S32 d0,d0,#0
+_NEON2SSESTORAGE uint64x1_t vqshlu_n_s64(int64x1_t a, __constrange(0,63) int b); // VQSHLU.S64 d0,d0,#0
+_NEON2SSESTORAGE uint8x16_t vqshluq_n_s8(int8x16_t a, __constrange(0,7) int b); // VQSHLU.S8 q0,q0,#0
+_NEON2SSESTORAGE uint16x8_t vqshluq_n_s16(int16x8_t a, __constrange(0,15) int b); // VQSHLU.S16 q0,q0,#0
+_NEON2SSESTORAGE uint32x4_t vqshluq_n_s32(int32x4_t a, __constrange(0,31) int b); // VQSHLU.S32 q0,q0,#0
+_NEON2SSESTORAGE uint64x2_t vqshluq_n_s64(int64x2_t a, __constrange(0,63) int b); // VQSHLU.S64 q0,q0,#0
 //Vector narrowing shift right by constant
-int8x8_t vshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VSHRN.I16 d0,q0,#8
-int16x4_t vshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VSHRN.I32 d0,q0,#16
-int32x2_t vshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VSHRN.I64 d0,q0,#32
-uint8x8_t vshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VSHRN.I16 d0,q0,#8
-uint16x4_t vshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VSHRN.I32 d0,q0,#16
-uint32x2_t vshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VSHRN.I64 d0,q0,#32
+_NEON2SSESTORAGE int8x8_t vshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VSHRN.I16 d0,q0,#8
+_NEON2SSESTORAGE int16x4_t vshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VSHRN.I32 d0,q0,#16
+_NEON2SSESTORAGE int32x2_t vshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VSHRN.I64 d0,q0,#32
+_NEON2SSESTORAGE uint8x8_t vshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VSHRN.I16 d0,q0,#8
+_NEON2SSESTORAGE uint16x4_t vshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VSHRN.I32 d0,q0,#16
+_NEON2SSESTORAGE uint32x2_t vshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VSHRN.I64 d0,q0,#32
 //Vector signed->unsigned narrowing saturating shift right by constant
-uint8x8_t vqshrun_n_s16(int16x8_t a, __constrange(1,8) int b); // VQSHRUN.S16 d0,q0,#8
-uint16x4_t vqshrun_n_s32(int32x4_t a, __constrange(1,16) int b); // VQSHRUN.S32 d0,q0,#16
-uint32x2_t vqshrun_n_s64(int64x2_t a, __constrange(1,32) int b); // VQSHRUN.S64 d0,q0,#32
+_NEON2SSESTORAGE uint8x8_t vqshrun_n_s16(int16x8_t a, __constrange(1,8) int b); // VQSHRUN.S16 d0,q0,#8
+_NEON2SSESTORAGE uint16x4_t vqshrun_n_s32(int32x4_t a, __constrange(1,16) int b); // VQSHRUN.S32 d0,q0,#16
+_NEON2SSESTORAGE uint32x2_t vqshrun_n_s64(int64x2_t a, __constrange(1,32) int b); // VQSHRUN.S64 d0,q0,#32
 //Vector signed->unsigned rounding narrowing saturating shift right by constant
-uint8x8_t vqrshrun_n_s16(int16x8_t a, __constrange(1,8) int b); // VQRSHRUN.S16 d0,q0,#8
-uint16x4_t vqrshrun_n_s32(int32x4_t a, __constrange(1,16) int b); // VQRSHRUN.S32 d0,q0,#16
-uint32x2_t vqrshrun_n_s64(int64x2_t a, __constrange(1,32) int b); // VQRSHRUN.S64 d0,q0,#32
+_NEON2SSESTORAGE uint8x8_t vqrshrun_n_s16(int16x8_t a, __constrange(1,8) int b); // VQRSHRUN.S16 d0,q0,#8
+_NEON2SSESTORAGE uint16x4_t vqrshrun_n_s32(int32x4_t a, __constrange(1,16) int b); // VQRSHRUN.S32 d0,q0,#16
+_NEON2SSESTORAGE uint32x2_t vqrshrun_n_s64(int64x2_t a, __constrange(1,32) int b); // VQRSHRUN.S64 d0,q0,#32
 //Vector narrowing saturating shift right by constant
-int8x8_t vqshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VQSHRN.S16 d0,q0,#8
-int16x4_t vqshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VQSHRN.S32 d0,q0,#16
-int32x2_t vqshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VQSHRN.S64 d0,q0,#32
-uint8x8_t vqshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VQSHRN.U16 d0,q0,#8
-uint16x4_t vqshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VQSHRN.U32 d0,q0,#16
-uint32x2_t vqshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VQSHRN.U64 d0,q0,#32
+_NEON2SSESTORAGE int8x8_t vqshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VQSHRN.S16 d0,q0,#8
+_NEON2SSESTORAGE int16x4_t vqshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VQSHRN.S32 d0,q0,#16
+_NEON2SSESTORAGE int32x2_t vqshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VQSHRN.S64 d0,q0,#32
+_NEON2SSESTORAGE uint8x8_t vqshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VQSHRN.U16 d0,q0,#8
+_NEON2SSESTORAGE uint16x4_t vqshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VQSHRN.U32 d0,q0,#16
+_NEON2SSESTORAGE uint32x2_t vqshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VQSHRN.U64 d0,q0,#32
 //Vector rounding narrowing shift right by constant
-int8x8_t vrshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VRSHRN.I16 d0,q0,#8
-int16x4_t vrshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VRSHRN.I32 d0,q0,#16
-int32x2_t vrshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VRSHRN.I64 d0,q0,#32
-uint8x8_t vrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VRSHRN.I16 d0,q0,#8
-uint16x4_t vrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VRSHRN.I32 d0,q0,#16
-uint32x2_t vrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VRSHRN.I64 d0,q0,#32
+_NEON2SSESTORAGE int8x8_t vrshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VRSHRN.I16 d0,q0,#8
+_NEON2SSESTORAGE int16x4_t vrshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VRSHRN.I32 d0,q0,#16
+_NEON2SSESTORAGE int32x2_t vrshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VRSHRN.I64 d0,q0,#32
+_NEON2SSESTORAGE uint8x8_t vrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VRSHRN.I16 d0,q0,#8
+_NEON2SSESTORAGE uint16x4_t vrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VRSHRN.I32 d0,q0,#16
+_NEON2SSESTORAGE uint32x2_t vrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VRSHRN.I64 d0,q0,#32
 //Vector rounding narrowing saturating shift right by constant
-int8x8_t vqrshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VQRSHRN.S16 d0,q0,#8
-int16x4_t vqrshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VQRSHRN.S32 d0,q0,#16
-int32x2_t vqrshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VQRSHRN.S64 d0,q0,#32
-uint8x8_t vqrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VQRSHRN.U16 d0,q0,#8
-uint16x4_t vqrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VQRSHRN.U32 d0,q0,#16
-uint32x2_t vqrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VQRSHRN.U64 d0,q0,#32
+_NEON2SSESTORAGE int8x8_t vqrshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VQRSHRN.S16 d0,q0,#8
+_NEON2SSESTORAGE int16x4_t vqrshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VQRSHRN.S32 d0,q0,#16
+_NEON2SSESTORAGE int32x2_t vqrshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VQRSHRN.S64 d0,q0,#32
+_NEON2SSESTORAGE uint8x8_t vqrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VQRSHRN.U16 d0,q0,#8
+_NEON2SSESTORAGE uint16x4_t vqrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VQRSHRN.U32 d0,q0,#16
+_NEON2SSESTORAGE uint32x2_t vqrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VQRSHRN.U64 d0,q0,#32
 //Vector widening shift left by constant
-int16x8_t vshll_n_s8(int8x8_t a, __constrange(0,8) int b); // VSHLL.S8 q0,d0,#0
-int32x4_t vshll_n_s16(int16x4_t a, __constrange(0,16) int b); // VSHLL.S16 q0,d0,#0
-int64x2_t vshll_n_s32(int32x2_t a, __constrange(0,32) int b); // VSHLL.S32 q0,d0,#0
-uint16x8_t vshll_n_u8(uint8x8_t a, __constrange(0,8) int b); // VSHLL.U8 q0,d0,#0
-uint32x4_t vshll_n_u16(uint16x4_t a, __constrange(0,16) int b); // VSHLL.U16 q0,d0,#0
-uint64x2_t vshll_n_u32(uint32x2_t a, __constrange(0,32) int b); // VSHLL.U32 q0,d0,#0
+_NEON2SSESTORAGE int16x8_t vshll_n_s8(int8x8_t a, __constrange(0,8) int b); // VSHLL.S8 q0,d0,#0
+_NEON2SSESTORAGE int32x4_t vshll_n_s16(int16x4_t a, __constrange(0,16) int b); // VSHLL.S16 q0,d0,#0
+_NEON2SSESTORAGE int64x2_t vshll_n_s32(int32x2_t a, __constrange(0,32) int b); // VSHLL.S32 q0,d0,#0
+_NEON2SSESTORAGE uint16x8_t vshll_n_u8(uint8x8_t a, __constrange(0,8) int b); // VSHLL.U8 q0,d0,#0
+_NEON2SSESTORAGE uint32x4_t vshll_n_u16(uint16x4_t a, __constrange(0,16) int b); // VSHLL.U16 q0,d0,#0
+_NEON2SSESTORAGE uint64x2_t vshll_n_u32(uint32x2_t a, __constrange(0,32) int b); // VSHLL.U32 q0,d0,#0
 //Shifts with insert
 //Vector shift right and insert
-int8x8_t vsri_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
-int16x4_t vsri_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
-int32x2_t vsri_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c); // VSRI.32 d0,d0,#32
-int64x1_t vsri_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VSRI.64 d0,d0,#64
-uint8x8_t vsri_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
-uint16x4_t vsri_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
-uint32x2_t vsri_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c); // VSRI.32 d0,d0,#32
-uint64x1_t vsri_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VSRI.64 d0,d0,#64
-poly8x8_t vsri_n_p8(poly8x8_t a, poly8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
-poly16x4_t vsri_n_p16(poly16x4_t a, poly16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
-int8x16_t vsriq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
-int16x8_t vsriq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
-int32x4_t vsriq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VSRI.32 q0,q0,#32
-int64x2_t vsriq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VSRI.64 q0,q0,#64
-uint8x16_t vsriq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
-uint16x8_t vsriq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
-uint32x4_t vsriq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VSRI.32 q0,q0,#32
-uint64x2_t vsriq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VSRI.64 q0,q0,#64
-poly8x16_t vsriq_n_p8(poly8x16_t a, poly8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
-poly16x8_t vsriq_n_p16(poly16x8_t a, poly16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
+_NEON2SSESTORAGE int8x8_t vsri_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
+_NEON2SSESTORAGE int16x4_t vsri_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
+_NEON2SSESTORAGE int32x2_t vsri_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c); // VSRI.32 d0,d0,#32
+_NEON2SSESTORAGE int64x1_t vsri_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VSRI.64 d0,d0,#64
+_NEON2SSESTORAGE uint8x8_t vsri_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
+_NEON2SSESTORAGE uint16x4_t vsri_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
+_NEON2SSESTORAGE uint32x2_t vsri_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c); // VSRI.32 d0,d0,#32
+_NEON2SSESTORAGE uint64x1_t vsri_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VSRI.64 d0,d0,#64
+_NEON2SSESTORAGE poly8x8_t vsri_n_p8(poly8x8_t a, poly8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
+_NEON2SSESTORAGE poly16x4_t vsri_n_p16(poly16x4_t a, poly16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
+_NEON2SSESTORAGE int8x16_t vsriq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
+_NEON2SSESTORAGE int16x8_t vsriq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
+_NEON2SSESTORAGE int32x4_t vsriq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VSRI.32 q0,q0,#32
+_NEON2SSESTORAGE int64x2_t vsriq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VSRI.64 q0,q0,#64
+_NEON2SSESTORAGE uint8x16_t vsriq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
+_NEON2SSESTORAGE uint16x8_t vsriq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
+_NEON2SSESTORAGE uint32x4_t vsriq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VSRI.32 q0,q0,#32
+_NEON2SSESTORAGE uint64x2_t vsriq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VSRI.64 q0,q0,#64
+_NEON2SSESTORAGE poly8x16_t vsriq_n_p8(poly8x16_t a, poly8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
+_NEON2SSESTORAGE poly16x8_t vsriq_n_p16(poly16x8_t a, poly16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
 //Vector shift left and insert
-int8x8_t vsli_n_s8(int8x8_t a, int8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
-int16x4_t vsli_n_s16(int16x4_t a, int16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
-int32x2_t vsli_n_s32(int32x2_t a, int32x2_t b, __constrange(0,31) int c); // VSLI.32 d0,d0,#0
-int64x1_t vsli_n_s64(int64x1_t a, int64x1_t b, __constrange(0,63) int c); // VSLI.64 d0,d0,#0
-uint8x8_t vsli_n_u8(uint8x8_t a, uint8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
-uint16x4_t vsli_n_u16(uint16x4_t a, uint16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
-uint32x2_t vsli_n_u32(uint32x2_t a, uint32x2_t b, __constrange(0,31) int c); // VSLI.32 d0,d0,#0
-uint64x1_t vsli_n_u64(uint64x1_t a, uint64x1_t b, __constrange(0,63) int c); // VSLI.64 d0,d0,#0
-poly8x8_t vsli_n_p8(poly8x8_t a, poly8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
-poly16x4_t vsli_n_p16(poly16x4_t a, poly16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
-int8x16_t vsliq_n_s8(int8x16_t a, int8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
-int16x8_t vsliq_n_s16(int16x8_t a, int16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
-int32x4_t vsliq_n_s32(int32x4_t a, int32x4_t b, __constrange(0,31) int c); // VSLI.32 q0,q0,#0
-int64x2_t vsliq_n_s64(int64x2_t a, int64x2_t b, __constrange(0,63) int c); // VSLI.64 q0,q0,#0
-uint8x16_t vsliq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
-uint16x8_t vsliq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
-uint32x4_t vsliq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(0,31) int c); // VSLI.32 q0,q0,#0
-uint64x2_t vsliq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(0,63) int c); // VSLI.64 q0,q0,#0
-poly8x16_t vsliq_n_p8(poly8x16_t a, poly8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
-poly16x8_t vsliq_n_p16(poly16x8_t a, poly16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
+_NEON2SSESTORAGE int8x8_t vsli_n_s8(int8x8_t a, int8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
+_NEON2SSESTORAGE int16x4_t vsli_n_s16(int16x4_t a, int16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
+_NEON2SSESTORAGE int32x2_t vsli_n_s32(int32x2_t a, int32x2_t b, __constrange(0,31) int c); // VSLI.32 d0,d0,#0
+_NEON2SSESTORAGE int64x1_t vsli_n_s64(int64x1_t a, int64x1_t b, __constrange(0,63) int c); // VSLI.64 d0,d0,#0
+_NEON2SSESTORAGE uint8x8_t vsli_n_u8(uint8x8_t a, uint8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
+_NEON2SSESTORAGE uint16x4_t vsli_n_u16(uint16x4_t a, uint16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
+_NEON2SSESTORAGE uint32x2_t vsli_n_u32(uint32x2_t a, uint32x2_t b, __constrange(0,31) int c); // VSLI.32 d0,d0,#0
+_NEON2SSESTORAGE uint64x1_t vsli_n_u64(uint64x1_t a, uint64x1_t b, __constrange(0,63) int c); // VSLI.64 d0,d0,#0
+_NEON2SSESTORAGE poly8x8_t vsli_n_p8(poly8x8_t a, poly8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
+_NEON2SSESTORAGE poly16x4_t vsli_n_p16(poly16x4_t a, poly16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
+_NEON2SSESTORAGE int8x16_t vsliq_n_s8(int8x16_t a, int8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
+_NEON2SSESTORAGE int16x8_t vsliq_n_s16(int16x8_t a, int16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
+_NEON2SSESTORAGE int32x4_t vsliq_n_s32(int32x4_t a, int32x4_t b, __constrange(0,31) int c); // VSLI.32 q0,q0,#0
+_NEON2SSESTORAGE int64x2_t vsliq_n_s64(int64x2_t a, int64x2_t b, __constrange(0,63) int c); // VSLI.64 q0,q0,#0
+_NEON2SSESTORAGE uint8x16_t vsliq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
+_NEON2SSESTORAGE uint16x8_t vsliq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
+_NEON2SSESTORAGE uint32x4_t vsliq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(0,31) int c); // VSLI.32 q0,q0,#0
+_NEON2SSESTORAGE uint64x2_t vsliq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(0,63) int c); // VSLI.64 q0,q0,#0
+_NEON2SSESTORAGE poly8x16_t vsliq_n_p8(poly8x16_t a, poly8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
+_NEON2SSESTORAGE poly16x8_t vsliq_n_p16(poly16x8_t a, poly16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
 //Loads of a single vector or lane. Perform loads and stores of a single vector of some type.
 //Load a single vector from memory
-uint8x16_t vld1q_u8(__transfersize(16) uint8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
-uint16x8_t vld1q_u16(__transfersize(8) uint16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
-uint32x4_t vld1q_u32(__transfersize(4) uint32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
-uint64x2_t vld1q_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
-int8x16_t vld1q_s8(__transfersize(16) int8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
-int16x8_t vld1q_s16(__transfersize(8) int16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
-int32x4_t vld1q_s32(__transfersize(4) int32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
-int64x2_t vld1q_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
-float16x8_t vld1q_f16(__transfersize(8) __fp16 const * ptr); // VLD1.16 {d0, d1}, [r0]
-float32x4_t vld1q_f32(__transfersize(4) float32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
-poly8x16_t vld1q_p8(__transfersize(16) poly8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
-poly16x8_t vld1q_p16(__transfersize(8) poly16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
-uint8x8_t vld1_u8(__transfersize(8) uint8_t const * ptr); // VLD1.8 {d0}, [r0]
-uint16x4_t vld1_u16(__transfersize(4) uint16_t const * ptr); // VLD1.16 {d0}, [r0]
-uint32x2_t vld1_u32(__transfersize(2) uint32_t const * ptr); // VLD1.32 {d0}, [r0]
-uint64x1_t vld1_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
-int8x8_t vld1_s8(__transfersize(8) int8_t const * ptr); // VLD1.8 {d0}, [r0]
-int16x4_t vld1_s16(__transfersize(4) int16_t const * ptr); // VLD1.16 {d0}, [r0]
-int32x2_t vld1_s32(__transfersize(2) int32_t const * ptr); // VLD1.32 {d0}, [r0]
-int64x1_t vld1_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
-float16x4_t vld1_f16(__transfersize(4) __fp16 const * ptr); // VLD1.16 {d0}, [r0]
-float32x2_t vld1_f32(__transfersize(2) float32_t const * ptr); // VLD1.32 {d0}, [r0]
-poly8x8_t vld1_p8(__transfersize(8) poly8_t const * ptr); // VLD1.8 {d0}, [r0]
-poly16x4_t vld1_p16(__transfersize(4) poly16_t const * ptr); // VLD1.16 {d0}, [r0]
-
-float64x2_t vld1q_f64(__transfersize(4) float64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint8x16_t vld1q_u8(__transfersize(16) uint8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint16x8_t vld1q_u16(__transfersize(8) uint16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint32x4_t vld1q_u32(__transfersize(4) uint32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint64x2_t vld1q_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE int8x16_t vld1q_s8(__transfersize(16) int8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE int16x8_t vld1q_s16(__transfersize(8) int16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE int32x4_t vld1q_s32(__transfersize(4) int32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE int64x2_t vld1q_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE float16x8_t vld1q_f16(__transfersize(8) __fp16 const * ptr); // VLD1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE float32x4_t vld1q_f32(__transfersize(4) float32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE poly8x16_t vld1q_p8(__transfersize(16) poly8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE poly16x8_t vld1q_p16(__transfersize(8) poly16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint8x8_t vld1_u8(__transfersize(8) uint8_t const * ptr); // VLD1.8 {d0}, [r0]
+_NEON2SSESTORAGE uint16x4_t vld1_u16(__transfersize(4) uint16_t const * ptr); // VLD1.16 {d0}, [r0]
+_NEON2SSESTORAGE uint32x2_t vld1_u32(__transfersize(2) uint32_t const * ptr); // VLD1.32 {d0}, [r0]
+_NEON2SSESTORAGE uint64x1_t vld1_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE int8x8_t vld1_s8(__transfersize(8) int8_t const * ptr); // VLD1.8 {d0}, [r0]
+_NEON2SSESTORAGE int16x4_t vld1_s16(__transfersize(4) int16_t const * ptr); // VLD1.16 {d0}, [r0]
+_NEON2SSESTORAGE int32x2_t vld1_s32(__transfersize(2) int32_t const * ptr); // VLD1.32 {d0}, [r0]
+_NEON2SSESTORAGE int64x1_t vld1_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE float16x4_t vld1_f16(__transfersize(4) __fp16 const * ptr); // VLD1.16 {d0}, [r0]
+_NEON2SSESTORAGE float32x2_t vld1_f32(__transfersize(2) float32_t const * ptr); // VLD1.32 {d0}, [r0]
+_NEON2SSESTORAGE poly8x8_t vld1_p8(__transfersize(8) poly8_t const * ptr); // VLD1.8 {d0}, [r0]
+_NEON2SSESTORAGE poly16x4_t vld1_p16(__transfersize(4) poly16_t const * ptr); // VLD1.16 {d0}, [r0]
+
+_NEON2SSESTORAGE float64x2_t vld1q_f64(__transfersize(4) float64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
 
 //Load a single lane from memory
-uint8x16_t vld1q_lane_u8(__transfersize(1) uint8_t const * ptr, uint8x16_t vec, __constrange(0,15) int lane); //VLD1.8 {d0[0]}, [r0]
-uint16x8_t vld1q_lane_u16(__transfersize(1) uint16_t const * ptr, uint16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
-uint32x4_t vld1q_lane_u32(__transfersize(1) uint32_t const * ptr, uint32x4_t vec, __constrange(0,3) int lane); // VLD1.32 {d0[0]}, [r0]
-uint64x2_t vld1q_lane_u64(__transfersize(1) uint64_t const * ptr, uint64x2_t vec, __constrange(0,1) int lane); // VLD1.64 {d0}, [r0]
-int8x16_t vld1q_lane_s8(__transfersize(1) int8_t const * ptr, int8x16_t vec, __constrange(0,15) int lane); //VLD1.8 {d0[0]}, [r0]
-int16x8_t vld1q_lane_s16(__transfersize(1) int16_t const * ptr, int16x8_t vec, __constrange(0,7) int lane); //VLD1.16 {d0[0]}, [r0]
-int32x4_t vld1q_lane_s32(__transfersize(1) int32_t const * ptr, int32x4_t vec, __constrange(0,3) int lane); //VLD1.32 {d0[0]}, [r0]
-float16x8_t vld1q_lane_f16(__transfersize(1) __fp16 const * ptr, float16x8_t vec, __constrange(0,7) int lane); //VLD1.16 {d0[0]}, [r0]
-float32x4_t vld1q_lane_f32(__transfersize(1) float32_t const * ptr, float32x4_t vec, __constrange(0,3) int lane); // VLD1.32 {d0[0]}, [r0]
-int64x2_t vld1q_lane_s64(__transfersize(1) int64_t const * ptr, int64x2_t vec, __constrange(0,1) int lane); //VLD1.64 {d0}, [r0]
-poly8x16_t vld1q_lane_p8(__transfersize(1) poly8_t const * ptr, poly8x16_t vec, __constrange(0,15) int lane); //VLD1.8 {d0[0]}, [r0]
-poly16x8_t vld1q_lane_p16(__transfersize(1) poly16_t const * ptr, poly16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
-uint8x8_t vld1_lane_u8(__transfersize(1) uint8_t const * ptr, uint8x8_t vec, __constrange(0,7) int lane); //VLD1.8 {d0[0]}, [r0]
-uint16x4_t vld1_lane_u16(__transfersize(1) uint16_t const * ptr, uint16x4_t vec, __constrange(0,3) int lane); //VLD1.16 {d0[0]}, [r0]
-uint32x2_t vld1_lane_u32(__transfersize(1) uint32_t const * ptr, uint32x2_t vec, __constrange(0,1) int lane); //VLD1.32 {d0[0]}, [r0]
-uint64x1_t vld1_lane_u64(__transfersize(1) uint64_t const * ptr, uint64x1_t vec, __constrange(0,0) int lane); //VLD1.64 {d0}, [r0]
-int8x8_t vld1_lane_s8(__transfersize(1) int8_t const * ptr, int8x8_t vec, __constrange(0,7) int lane); // VLD1.8{d0[0]}, [r0]
-int16x4_t vld1_lane_s16(__transfersize(1) int16_t const * ptr, int16x4_t vec, __constrange(0,3) int lane); //VLD1.16 {d0[0]}, [r0]
-int32x2_t vld1_lane_s32(__transfersize(1) int32_t const * ptr, int32x2_t vec, __constrange(0,1) int lane); //VLD1.32 {d0[0]}, [r0]
-float16x4_t vld1q_lane_f16(__transfersize(1) __fp16 const * ptr, float16x4_t vec, __constrange(0,3) int lane); //VLD1.16 {d0[0]}, [r0]
-float32x2_t vld1_lane_f32(__transfersize(1) float32_t const * ptr, float32x2_t vec, __constrange(0,1) int lane); // VLD1.32 {d0[0]}, [r0]
-int64x1_t vld1_lane_s64(__transfersize(1) int64_t const * ptr, int64x1_t vec, __constrange(0,0) int lane); //VLD1.64 {d0}, [r0]
-poly8x8_t vld1_lane_p8(__transfersize(1) poly8_t const * ptr, poly8x8_t vec, __constrange(0,7) int lane); //VLD1.8 {d0[0]}, [r0]
-poly16x4_t vld1_lane_p16(__transfersize(1) poly16_t const * ptr, poly16x4_t vec, __constrange(0,3) int lane); //VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint8x16_t vld1q_lane_u8(__transfersize(1) uint8_t const * ptr, uint8x16_t vec, __constrange(0,15) int lane); //VLD1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint16x8_t vld1q_lane_u16(__transfersize(1) uint16_t const * ptr, uint16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint32x4_t vld1q_lane_u32(__transfersize(1) uint32_t const * ptr, uint32x4_t vec, __constrange(0,3) int lane); // VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint64x2_t vld1q_lane_u64(__transfersize(1) uint64_t const * ptr, uint64x2_t vec, __constrange(0,1) int lane); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE int8x16_t vld1q_lane_s8(__transfersize(1) int8_t const * ptr, int8x16_t vec, __constrange(0,15) int lane); //VLD1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE int16x8_t vld1q_lane_s16(__transfersize(1) int16_t const * ptr, int16x8_t vec, __constrange(0,7) int lane); //VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE int32x4_t vld1q_lane_s32(__transfersize(1) int32_t const * ptr, int32x4_t vec, __constrange(0,3) int lane); //VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE float16x8_t vld1q_lane_f16(__transfersize(1) __fp16 const * ptr, float16x8_t vec, __constrange(0,7) int lane); //VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE float32x4_t vld1q_lane_f32(__transfersize(1) float32_t const * ptr, float32x4_t vec, __constrange(0,3) int lane); // VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE int64x2_t vld1q_lane_s64(__transfersize(1) int64_t const * ptr, int64x2_t vec, __constrange(0,1) int lane); //VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE poly8x16_t vld1q_lane_p8(__transfersize(1) poly8_t const * ptr, poly8x16_t vec, __constrange(0,15) int lane); //VLD1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE poly16x8_t vld1q_lane_p16(__transfersize(1) poly16_t const * ptr, poly16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint8x8_t vld1_lane_u8(__transfersize(1) uint8_t const * ptr, uint8x8_t vec, __constrange(0,7) int lane); //VLD1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint16x4_t vld1_lane_u16(__transfersize(1) uint16_t const * ptr, uint16x4_t vec, __constrange(0,3) int lane); //VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint32x2_t vld1_lane_u32(__transfersize(1) uint32_t const * ptr, uint32x2_t vec, __constrange(0,1) int lane); //VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint64x1_t vld1_lane_u64(__transfersize(1) uint64_t const * ptr, uint64x1_t vec, __constrange(0,0) int lane); //VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE int8x8_t vld1_lane_s8(__transfersize(1) int8_t const * ptr, int8x8_t vec, __constrange(0,7) int lane); // VLD1.8{d0[0]}, [r0]
+_NEON2SSESTORAGE int16x4_t vld1_lane_s16(__transfersize(1) int16_t const * ptr, int16x4_t vec, __constrange(0,3) int lane); //VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE int32x2_t vld1_lane_s32(__transfersize(1) int32_t const * ptr, int32x2_t vec, __constrange(0,1) int lane); //VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE float16x4_t vld1q_lane_f16(__transfersize(1) __fp16 const * ptr, float16x4_t vec, __constrange(0,3) int lane); //VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE float32x2_t vld1_lane_f32(__transfersize(1) float32_t const * ptr, float32x2_t vec, __constrange(0,1) int lane); // VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE int64x1_t vld1_lane_s64(__transfersize(1) int64_t const * ptr, int64x1_t vec, __constrange(0,0) int lane); //VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE poly8x8_t vld1_lane_p8(__transfersize(1) poly8_t const * ptr, poly8x8_t vec, __constrange(0,7) int lane); //VLD1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE poly16x4_t vld1_lane_p16(__transfersize(1) poly16_t const * ptr, poly16x4_t vec, __constrange(0,3) int lane); //VLD1.16 {d0[0]}, [r0]
 //Load all lanes of vector with same value from memory
-uint8x16_t vld1q_dup_u8(__transfersize(1) uint8_t const * ptr); // VLD1.8 {d0[]}, [r0]
-uint16x8_t vld1q_dup_u16(__transfersize(1) uint16_t const * ptr); // VLD1.16 {d0[]}, [r0]
-uint32x4_t vld1q_dup_u32(__transfersize(1) uint32_t const * ptr); // VLD1.32 {d0[]}, [r0]
-uint64x2_t vld1q_dup_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
-int8x16_t vld1q_dup_s8(__transfersize(1) int8_t const * ptr); // VLD1.8 {d0[]}, [r0]
-int16x8_t vld1q_dup_s16(__transfersize(1) int16_t const * ptr); // VLD1.16 {d0[]}, [r0]
-int32x4_t vld1q_dup_s32(__transfersize(1) int32_t const * ptr); // VLD1.32 {d0[]}, [r0]
-int64x2_t vld1q_dup_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
-float16x8_t vld1q_dup_f16(__transfersize(1) __fp16 const * ptr); // VLD1.16 {d0[]}, [r0]
-float32x4_t vld1q_dup_f32(__transfersize(1) float32_t const * ptr); // VLD1.32 {d0[]}, [r0]
-poly8x16_t vld1q_dup_p8(__transfersize(1) poly8_t const * ptr); // VLD1.8 {d0[]}, [r0]
-poly16x8_t vld1q_dup_p16(__transfersize(1) poly16_t const * ptr); // VLD1.16 {d0[]}, [r0]
-uint8x8_t vld1_dup_u8(__transfersize(1) uint8_t const * ptr); // VLD1.8 {d0[]}, [r0]
-uint16x4_t vld1_dup_u16(__transfersize(1) uint16_t const * ptr); // VLD1.16 {d0[]}, [r0]
-uint32x2_t vld1_dup_u32(__transfersize(1) uint32_t const * ptr); // VLD1.32 {d0[]}, [r0]
-uint64x1_t vld1_dup_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
-int8x8_t vld1_dup_s8(__transfersize(1) int8_t const * ptr); // VLD1.8 {d0[]}, [r0]
-int16x4_t vld1_dup_s16(__transfersize(1) int16_t const * ptr); // VLD1.16 {d0[]}, [r0]
-int32x2_t vld1_dup_s32(__transfersize(1) int32_t const * ptr); // VLD1.32 {d0[]}, [r0]
-int64x1_t vld1_dup_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
-float16x4_t vld1_dup_f16(__transfersize(1) __fp16 const * ptr); // VLD1.16 {d0[]}, [r0]
-float32x2_t vld1_dup_f32(__transfersize(1) float32_t const * ptr); // VLD1.32 {d0[]}, [r0]
-poly8x8_t vld1_dup_p8(__transfersize(1) poly8_t const * ptr); // VLD1.8 {d0[]}, [r0]
-poly16x4_t vld1_dup_p16(__transfersize(1) poly16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE uint8x16_t vld1q_dup_u8(__transfersize(1) uint8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE uint16x8_t vld1q_dup_u16(__transfersize(1) uint16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE uint32x4_t vld1q_dup_u32(__transfersize(1) uint32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE uint64x2_t vld1q_dup_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE int8x16_t vld1q_dup_s8(__transfersize(1) int8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE int16x8_t vld1q_dup_s16(__transfersize(1) int16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE int32x4_t vld1q_dup_s32(__transfersize(1) int32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE int64x2_t vld1q_dup_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE float16x8_t vld1q_dup_f16(__transfersize(1) __fp16 const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE float32x4_t vld1q_dup_f32(__transfersize(1) float32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE poly8x16_t vld1q_dup_p8(__transfersize(1) poly8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE poly16x8_t vld1q_dup_p16(__transfersize(1) poly16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE uint8x8_t vld1_dup_u8(__transfersize(1) uint8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE uint16x4_t vld1_dup_u16(__transfersize(1) uint16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE uint32x2_t vld1_dup_u32(__transfersize(1) uint32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE uint64x1_t vld1_dup_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE int8x8_t vld1_dup_s8(__transfersize(1) int8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE int16x4_t vld1_dup_s16(__transfersize(1) int16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE int32x2_t vld1_dup_s32(__transfersize(1) int32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE int64x1_t vld1_dup_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE float16x4_t vld1_dup_f16(__transfersize(1) __fp16 const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE float32x2_t vld1_dup_f32(__transfersize(1) float32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE poly8x8_t vld1_dup_p8(__transfersize(1) poly8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE poly16x4_t vld1_dup_p16(__transfersize(1) poly16_t const * ptr); // VLD1.16 {d0[]}, [r0]
 //Store a single vector or lane. Stores all lanes or a single lane of a vector.
 //Store a single vector into memory
-void vst1q_u8(__transfersize(16) uint8_t * ptr, uint8x16_t val); // VST1.8 {d0, d1}, [r0]
-void vst1q_u16(__transfersize(8) uint16_t * ptr, uint16x8_t val); // VST1.16 {d0, d1}, [r0]
-void vst1q_u32(__transfersize(4) uint32_t * ptr, uint32x4_t val); // VST1.32 {d0, d1}, [r0]
-void vst1q_u64(__transfersize(2) uint64_t * ptr, uint64x2_t val); // VST1.64 {d0, d1}, [r0]
-void vst1q_s8(__transfersize(16) int8_t * ptr, int8x16_t val); // VST1.8 {d0, d1}, [r0]
-void vst1q_s16(__transfersize(8) int16_t * ptr, int16x8_t val); // VST1.16 {d0, d1}, [r0]
-void vst1q_s32(__transfersize(4) int32_t * ptr, int32x4_t val); // VST1.32 {d0, d1}, [r0]
-void vst1q_s64(__transfersize(2) int64_t * ptr, int64x2_t val); // VST1.64 {d0, d1}, [r0]
-void vst1q_f16(__transfersize(8) __fp16 * ptr, float16x8_t val); // VST1.16 {d0, d1}, [r0]
-void vst1q_f32(__transfersize(4) float32_t * ptr, float32x4_t val); // VST1.32 {d0, d1}, [r0]
-void vst1q_p8(__transfersize(16) poly8_t * ptr, poly8x16_t val); // VST1.8 {d0, d1}, [r0]
-void vst1q_p16(__transfersize(8) poly16_t * ptr, poly16x8_t val); // VST1.16 {d0, d1}, [r0]
-void vst1_u8(__transfersize(8) uint8_t * ptr, uint8x8_t val); // VST1.8 {d0}, [r0]
-void vst1_u16(__transfersize(4) uint16_t * ptr, uint16x4_t val); // VST1.16 {d0}, [r0]
-void vst1_u32(__transfersize(2) uint32_t * ptr, uint32x2_t val); // VST1.32 {d0}, [r0]
-void vst1_u64(__transfersize(1) uint64_t * ptr, uint64x1_t val); // VST1.64 {d0}, [r0]
-void vst1_s8(__transfersize(8) int8_t * ptr, int8x8_t val); // VST1.8 {d0}, [r0]
-void vst1_s16(__transfersize(4) int16_t * ptr, int16x4_t val); // VST1.16 {d0}, [r0]
-void vst1_s32(__transfersize(2) int32_t * ptr, int32x2_t val); // VST1.32 {d0}, [r0]
-void vst1_s64(__transfersize(1) int64_t * ptr, int64x1_t val); // VST1.64 {d0}, [r0]
-void vst1_f16(__transfersize(4) __fp16 * ptr, float16x4_t val); // VST1.16 {d0}, [r0]
-void vst1_f32(__transfersize(2) float32_t * ptr, float32x2_t val); // VST1.32 {d0}, [r0]
-void vst1_p8(__transfersize(8) poly8_t * ptr, poly8x8_t val); // VST1.8 {d0}, [r0]
-void vst1_p16(__transfersize(4) poly16_t * ptr, poly16x4_t val); // VST1.16 {d0}, [r0]
+_NEON2SSESTORAGE void vst1q_u8(__transfersize(16) uint8_t * ptr, uint8x16_t val); // VST1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_u16(__transfersize(8) uint16_t * ptr, uint16x8_t val); // VST1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_u32(__transfersize(4) uint32_t * ptr, uint32x4_t val); // VST1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_u64(__transfersize(2) uint64_t * ptr, uint64x2_t val); // VST1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_s8(__transfersize(16) int8_t * ptr, int8x16_t val); // VST1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_s16(__transfersize(8) int16_t * ptr, int16x8_t val); // VST1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_s32(__transfersize(4) int32_t * ptr, int32x4_t val); // VST1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_s64(__transfersize(2) int64_t * ptr, int64x2_t val); // VST1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_f16(__transfersize(8) __fp16 * ptr, float16x8_t val); // VST1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_f32(__transfersize(4) float32_t * ptr, float32x4_t val); // VST1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_p8(__transfersize(16) poly8_t * ptr, poly8x16_t val); // VST1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_p16(__transfersize(8) poly16_t * ptr, poly16x8_t val); // VST1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1_u8(__transfersize(8) uint8_t * ptr, uint8x8_t val); // VST1.8 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_u16(__transfersize(4) uint16_t * ptr, uint16x4_t val); // VST1.16 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_u32(__transfersize(2) uint32_t * ptr, uint32x2_t val); // VST1.32 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_u64(__transfersize(1) uint64_t * ptr, uint64x1_t val); // VST1.64 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_s8(__transfersize(8) int8_t * ptr, int8x8_t val); // VST1.8 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_s16(__transfersize(4) int16_t * ptr, int16x4_t val); // VST1.16 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_s32(__transfersize(2) int32_t * ptr, int32x2_t val); // VST1.32 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_s64(__transfersize(1) int64_t * ptr, int64x1_t val); // VST1.64 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_f16(__transfersize(4) __fp16 * ptr, float16x4_t val); // VST1.16 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_f32(__transfersize(2) float32_t * ptr, float32x2_t val); // VST1.32 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_p8(__transfersize(8) poly8_t * ptr, poly8x8_t val); // VST1.8 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_p16(__transfersize(4) poly16_t * ptr, poly16x4_t val); // VST1.16 {d0}, [r0]
 //Store a lane of a vector into memory
 //Loads of an N-element structure
 //Load N-element structure from memory
-uint8x16x2_t vld2q_u8(__transfersize(32) uint8_t const * ptr); // VLD2.8 {d0, d2}, [r0]
-uint16x8x2_t vld2q_u16(__transfersize(16) uint16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
-uint32x4x2_t vld2q_u32(__transfersize(8) uint32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
-int8x16x2_t vld2q_s8(__transfersize(32) int8_t const * ptr); // VLD2.8 {d0, d2}, [r0]
-int16x8x2_t vld2q_s16(__transfersize(16) int16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
-int32x4x2_t vld2q_s32(__transfersize(8) int32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
-float16x8x2_t vld2q_f16(__transfersize(16) __fp16 const * ptr); // VLD2.16 {d0, d2}, [r0]
-float32x4x2_t vld2q_f32(__transfersize(8) float32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
-poly8x16x2_t vld2q_p8(__transfersize(32) poly8_t const * ptr); // VLD2.8 {d0, d2}, [r0]
-poly16x8x2_t vld2q_p16(__transfersize(16) poly16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
-uint8x8x2_t vld2_u8(__transfersize(16) uint8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
-uint16x4x2_t vld2_u16(__transfersize(8) uint16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
-uint32x2x2_t vld2_u32(__transfersize(4) uint32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
-uint64x1x2_t vld2_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
-int8x8x2_t vld2_s8(__transfersize(16) int8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
-int16x4x2_t vld2_s16(__transfersize(8) int16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
-int32x2x2_t vld2_s32(__transfersize(4) int32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
-int64x1x2_t vld2_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint8x16x2_t vld2q_u8(__transfersize(32) uint8_t const * ptr); // VLD2.8 {d0, d2}, [r0]
+_NEON2SSESTORAGE uint16x8x2_t vld2q_u16(__transfersize(16) uint16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE uint32x4x2_t vld2q_u32(__transfersize(8) uint32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
+_NEON2SSESTORAGE int8x16x2_t vld2q_s8(__transfersize(32) int8_t const * ptr); // VLD2.8 {d0, d2}, [r0]
+_NEON2SSESTORAGE int16x8x2_t vld2q_s16(__transfersize(16) int16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE int32x4x2_t vld2q_s32(__transfersize(8) int32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
+_NEON2SSESTORAGE float16x8x2_t vld2q_f16(__transfersize(16) __fp16 const * ptr); // VLD2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE float32x4x2_t vld2q_f32(__transfersize(8) float32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
+_NEON2SSESTORAGE poly8x16x2_t vld2q_p8(__transfersize(32) poly8_t const * ptr); // VLD2.8 {d0, d2}, [r0]
+_NEON2SSESTORAGE poly16x8x2_t vld2q_p16(__transfersize(16) poly16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE uint8x8x2_t vld2_u8(__transfersize(16) uint8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint16x4x2_t vld2_u16(__transfersize(8) uint16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint32x2x2_t vld2_u32(__transfersize(4) uint32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint64x1x2_t vld2_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE int8x8x2_t vld2_s8(__transfersize(16) int8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE int16x4x2_t vld2_s16(__transfersize(8) int16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE int32x2x2_t vld2_s32(__transfersize(4) int32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE int64x1x2_t vld2_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
 //float16x4x2_t vld2_f16(__transfersize(8) __fp16 const * ptr); // VLD2.16 {d0, d1}, [r0]
-float32x2x2_t vld2_f32(__transfersize(4) float32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
-poly8x8x2_t vld2_p8(__transfersize(16) poly8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
-poly16x4x2_t vld2_p16(__transfersize(8) poly16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
-uint8x16x3_t vld3q_u8(__transfersize(48) uint8_t const * ptr); // VLD3.8 {d0, d2, d4}, [r0]
-uint16x8x3_t vld3q_u16(__transfersize(24) uint16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
-uint32x4x3_t vld3q_u32(__transfersize(12) uint32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
-int8x16x3_t vld3q_s8(__transfersize(48) int8_t const * ptr); // VLD3.8 {d0, d2, d4}, [r0]
-int16x8x3_t vld3q_s16(__transfersize(24) int16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
-int32x4x3_t vld3q_s32(__transfersize(12) int32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
-float16x8x3_t vld3q_f16(__transfersize(24) __fp16 const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
-float32x4x3_t vld3q_f32(__transfersize(12) float32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE float32x2x2_t vld2_f32(__transfersize(4) float32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE poly8x8x2_t vld2_p8(__transfersize(16) poly8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE poly16x4x2_t vld2_p16(__transfersize(8) poly16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint8x16x3_t vld3q_u8(__transfersize(48) uint8_t const * ptr); // VLD3.8 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE uint16x8x3_t vld3q_u16(__transfersize(24) uint16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE uint32x4x3_t vld3q_u32(__transfersize(12) uint32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE int8x16x3_t vld3q_s8(__transfersize(48) int8_t const * ptr); // VLD3.8 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE int16x8x3_t vld3q_s16(__transfersize(24) int16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE int32x4x3_t vld3q_s32(__transfersize(12) int32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE float16x8x3_t vld3q_f16(__transfersize(24) __fp16 const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE float32x4x3_t vld3q_f32(__transfersize(12) float32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
 poly8x16x3_t vld3q_p8(__transfersize(48) poly8_t const * ptr); // VLD3.8 {d0, d2, d4}, [r0]
-poly16x8x3_t vld3q_p16(__transfersize(24) poly16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
-uint8x8x3_t vld3_u8(__transfersize(24) uint8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
-uint16x4x3_t vld3_u16(__transfersize(12) uint16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
-uint32x2x3_t vld3_u32(__transfersize(6) uint32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
-uint64x1x3_t vld3_u64(__transfersize(3) uint64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
-int8x8x3_t vld3_s8(__transfersize(24) int8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
-int16x4x3_t vld3_s16(__transfersize(12) int16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
-int32x2x3_t vld3_s32(__transfersize(6) int32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE poly16x8x3_t vld3q_p16(__transfersize(24) poly16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE uint8x8x3_t vld3_u8(__transfersize(24) uint8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE uint16x4x3_t vld3_u16(__transfersize(12) uint16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE uint32x2x3_t vld3_u32(__transfersize(6) uint32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE uint64x1x3_t vld3_u64(__transfersize(3) uint64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE int8x8x3_t vld3_s8(__transfersize(24) int8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE int16x4x3_t vld3_s16(__transfersize(12) int16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE int32x2x3_t vld3_s32(__transfersize(6) int32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
 int64x1x3_t vld3_s64(__transfersize(3) int64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
-float16x4x3_t vld3_f16(__transfersize(12) __fp16 const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
-float32x2x3_t vld3_f32(__transfersize(6) float32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
-poly8x8x3_t vld3_p8(__transfersize(24) poly8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
-poly16x4x3_t vld3_p16(__transfersize(12) poly16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
-uint8x16x4_t vld4q_u8(__transfersize(64) uint8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
-uint16x8x4_t vld4q_u16(__transfersize(32) uint16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
-uint32x4x4_t vld4q_u32(__transfersize(16) uint32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
-int8x16x4_t vld4q_s8(__transfersize(64) int8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
-int16x8x4_t vld4q_s16(__transfersize(32) int16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
-int32x4x4_t vld4q_s32(__transfersize(16) int32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
-float16x8x4_t vld4q_f16(__transfersize(32) __fp16 const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
-float32x4x4_t vld4q_f32(__transfersize(16) float32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
-poly8x16x4_t vld4q_p8(__transfersize(64) poly8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
-poly16x8x4_t vld4q_p16(__transfersize(32) poly16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
-uint8x8x4_t vld4_u8(__transfersize(32) uint8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
-uint16x4x4_t vld4_u16(__transfersize(16) uint16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
-uint32x2x4_t vld4_u32(__transfersize(8) uint32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
-uint64x1x4_t vld4_u64(__transfersize(4) uint64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
-int8x8x4_t vld4_s8(__transfersize(32) int8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
-int16x4x4_t vld4_s16(__transfersize(16) int16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
-int32x2x4_t vld4_s32(__transfersize(8) int32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE float16x4x3_t vld3_f16(__transfersize(12) __fp16 const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE float32x2x3_t vld3_f32(__transfersize(6) float32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE poly8x8x3_t vld3_p8(__transfersize(24) poly8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE poly16x4x3_t vld3_p16(__transfersize(12) poly16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE uint8x16x4_t vld4q_u8(__transfersize(64) uint8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE uint16x8x4_t vld4q_u16(__transfersize(32) uint16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE uint32x4x4_t vld4q_u32(__transfersize(16) uint32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE int8x16x4_t vld4q_s8(__transfersize(64) int8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE int16x8x4_t vld4q_s16(__transfersize(32) int16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE int32x4x4_t vld4q_s32(__transfersize(16) int32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE float16x8x4_t vld4q_f16(__transfersize(32) __fp16 const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE float32x4x4_t vld4q_f32(__transfersize(16) float32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE poly8x16x4_t vld4q_p8(__transfersize(64) poly8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE poly16x8x4_t vld4q_p16(__transfersize(32) poly16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE uint8x8x4_t vld4_u8(__transfersize(32) uint8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE uint16x4x4_t vld4_u16(__transfersize(16) uint16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE uint32x2x4_t vld4_u32(__transfersize(8) uint32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE uint64x1x4_t vld4_u64(__transfersize(4) uint64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE int8x8x4_t vld4_s8(__transfersize(32) int8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE int16x4x4_t vld4_s16(__transfersize(16) int16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE int32x2x4_t vld4_s32(__transfersize(8) int32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
 int64x1x4_t vld4_s64(__transfersize(4) int64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
-float16x4x4_t vld4_f16(__transfersize(16) __fp16 const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
-float32x2x4_t vld4_f32(__transfersize(8) float32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
-poly8x8x4_t vld4_p8(__transfersize(32) poly8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
-poly16x4x4_t vld4_p16(__transfersize(16) poly16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE float16x4x4_t vld4_f16(__transfersize(16) __fp16 const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE float32x2x4_t vld4_f32(__transfersize(8) float32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE poly8x8x4_t vld4_p8(__transfersize(32) poly8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE poly16x4x4_t vld4_p16(__transfersize(16) poly16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
 //Load all lanes of N-element structure with same value from memory
-uint8x8x2_t vld2_dup_u8(__transfersize(2) uint8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
-uint16x4x2_t vld2_dup_u16(__transfersize(2) uint16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
-uint32x2x2_t vld2_dup_u32(__transfersize(2) uint32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
-uint64x1x2_t vld2_dup_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
-int8x8x2_t vld2_dup_s8(__transfersize(2) int8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
-int16x4x2_t vld2_dup_s16(__transfersize(2) int16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
-int32x2x2_t vld2_dup_s32(__transfersize(2) int32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
-int64x1x2_t vld2_dup_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint8x8x2_t vld2_dup_u8(__transfersize(2) uint8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE uint16x4x2_t vld2_dup_u16(__transfersize(2) uint16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE uint32x2x2_t vld2_dup_u32(__transfersize(2) uint32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE uint64x1x2_t vld2_dup_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE int8x8x2_t vld2_dup_s8(__transfersize(2) int8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE int16x4x2_t vld2_dup_s16(__transfersize(2) int16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE int32x2x2_t vld2_dup_s32(__transfersize(2) int32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE int64x1x2_t vld2_dup_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
 //float16x4x2_t vld2_dup_f16(__transfersize(2) __fp16 const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
-float32x2x2_t vld2_dup_f32(__transfersize(2) float32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
-poly8x8x2_t vld2_dup_p8(__transfersize(2) poly8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
-poly16x4x2_t vld2_dup_p16(__transfersize(2) poly16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
-uint8x8x3_t vld3_dup_u8(__transfersize(3) uint8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
-uint16x4x3_t vld3_dup_u16(__transfersize(3) uint16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
-uint32x2x3_t vld3_dup_u32(__transfersize(3) uint32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
-uint64x1x3_t vld3_dup_u64(__transfersize(3) uint64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
-int8x8x3_t vld3_dup_s8(__transfersize(3) int8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
-int16x4x3_t vld3_dup_s16(__transfersize(3) int16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
-int32x2x3_t vld3_dup_s32(__transfersize(3) int32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE float32x2x2_t vld2_dup_f32(__transfersize(2) float32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE poly8x8x2_t vld2_dup_p8(__transfersize(2) poly8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE poly16x4x2_t vld2_dup_p16(__transfersize(2) poly16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE uint8x8x3_t vld3_dup_u8(__transfersize(3) uint8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE uint16x4x3_t vld3_dup_u16(__transfersize(3) uint16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE uint32x2x3_t vld3_dup_u32(__transfersize(3) uint32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE uint64x1x3_t vld3_dup_u64(__transfersize(3) uint64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE int8x8x3_t vld3_dup_s8(__transfersize(3) int8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE int16x4x3_t vld3_dup_s16(__transfersize(3) int16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE int32x2x3_t vld3_dup_s32(__transfersize(3) int32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
 int64x1x3_t vld3_dup_s64(__transfersize(3) int64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
-float16x4x3_t vld3_dup_f16(__transfersize(3) __fp16 const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
-float32x2x3_t vld3_dup_f32(__transfersize(3) float32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
-poly8x8x3_t vld3_dup_p8(__transfersize(3) poly8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
-poly16x4x3_t vld3_dup_p16(__transfersize(3) poly16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
-uint8x8x4_t vld4_dup_u8(__transfersize(4) uint8_t const * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
-uint16x4x4_t vld4_dup_u16(__transfersize(4) uint16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
-uint32x2x4_t vld4_dup_u32(__transfersize(4) uint32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
-uint64x1x4_t vld4_dup_u64(__transfersize(4) uint64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
-int8x8x4_t vld4_dup_s8(__transfersize(4) int8_t const * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
-int16x4x4_t vld4_dup_s16(__transfersize(4) int16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
-int32x2x4_t vld4_dup_s32(__transfersize(4) int32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE float16x4x3_t vld3_dup_f16(__transfersize(3) __fp16 const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE float32x2x3_t vld3_dup_f32(__transfersize(3) float32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE poly8x8x3_t vld3_dup_p8(__transfersize(3) poly8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE poly16x4x3_t vld3_dup_p16(__transfersize(3) poly16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE uint8x8x4_t vld4_dup_u8(__transfersize(4) uint8_t const * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE uint16x4x4_t vld4_dup_u16(__transfersize(4) uint16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE uint32x2x4_t vld4_dup_u32(__transfersize(4) uint32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE uint64x1x4_t vld4_dup_u64(__transfersize(4) uint64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE int8x8x4_t vld4_dup_s8(__transfersize(4) int8_t const * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE int16x4x4_t vld4_dup_s16(__transfersize(4) int16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE int32x2x4_t vld4_dup_s32(__transfersize(4) int32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
 int64x1x4_t vld4_dup_s64(__transfersize(4) int64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
-float16x4x4_t vld4_dup_f16(__transfersize(4) __fp16 const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
-float32x2x4_t vld4_dup_f32(__transfersize(4) float32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
-poly8x8x4_t vld4_dup_p8(__transfersize(4) poly8_t const * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
-poly16x4x4_t vld4_dup_p16(__transfersize(4) poly16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE float16x4x4_t vld4_dup_f16(__transfersize(4) __fp16 const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE float32x2x4_t vld4_dup_f32(__transfersize(4) float32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE poly8x8x4_t vld4_dup_p8(__transfersize(4) poly8_t const * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE poly16x4x4_t vld4_dup_p16(__transfersize(4) poly16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
 //Load a single lane of N-element structure from memory
 //the functions below are modified to deal with the error C2719: 'src': formal parameter with __declspec(align('16')) won't be aligned
-uint16x8x2_t vld2q_lane_u16_ptr(__transfersize(2) uint16_t const * ptr, uint16x8x2_t * src, __constrange(0,7) int lane); // VLD2.16 {d0[0], d2[0]}, [r0]
-uint32x4x2_t vld2q_lane_u32_ptr(__transfersize(2) uint32_t const * ptr, uint32x4x2_t * src, __constrange(0,3) int lane); // VLD2.32 {d0[0], d2[0]}, [r0]
-int16x8x2_t vld2q_lane_s16_ptr(__transfersize(2) int16_t const * ptr, int16x8x2_t * src, __constrange(0,7) int lane); // VLD2.16 {d0[0], d2[0]}, [r0]
-int32x4x2_t vld2q_lane_s32_ptr(__transfersize(2) int32_t const * ptr, int32x4x2_t * src, __constrange(0,3) int lane); // VLD2.32 {d0[0], d2[0]}, [r0]
-float16x8x2_t vld2q_lane_f16_ptr(__transfersize(2) __fp16 const * ptr, float16x8x2_t * src, __constrange(0,7) int lane); // VLD2.16 {d0[0], d2[0]}, [r0]
-float32x4x2_t vld2q_lane_f32_ptr(__transfersize(2) float32_t const * ptr, float32x4x2_t * src, __constrange(0,3) int lane); // VLD2.32 {d0[0], d2[0]}, [r0]
-poly16x8x2_t vld2q_lane_p16_ptr(__transfersize(2) poly16_t const * ptr, poly16x8x2_t * src, __constrange(0,7) int lane); // VLD2.16 {d0[0], d2[0]}, [r0]
-uint8x8x2_t vld2_lane_u8(__transfersize(2) uint8_t const * ptr, uint8x8x2_t src, __constrange(0,7) int lane); //VLD2.8 {d0[0], d1[0]}, [r0]
-uint16x4x2_t vld2_lane_u16(__transfersize(2) uint16_t const * ptr, uint16x4x2_t src, __constrange(0,3) int lane); // VLD2.16 {d0[0], d1[0]}, [r0]
-uint32x2x2_t vld2_lane_u32(__transfersize(2) uint32_t const * ptr, uint32x2x2_t src, __constrange(0,1) int lane); // VLD2.32 {d0[0], d1[0]}, [r0]
-int8x8x2_t vld2_lane_s8(__transfersize(2) int8_t const * ptr, int8x8x2_t src, __constrange(0,7) int lane); //VLD2.8 {d0[0], d1[0]}, [r0]
-int16x4x2_t vld2_lane_s16(__transfersize(2) int16_t const * ptr, int16x4x2_t src, __constrange(0,3) int lane); //VLD2.16 {d0[0], d1[0]}, [r0]
-int32x2x2_t vld2_lane_s32(__transfersize(2) int32_t const * ptr, int32x2x2_t src, __constrange(0,1) int lane); //VLD2.32 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE uint16x8x2_t vld2q_lane_u16_ptr(__transfersize(2) uint16_t const * ptr, uint16x8x2_t * src, __constrange(0,7) int lane); // VLD2.16 {d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE uint32x4x2_t vld2q_lane_u32_ptr(__transfersize(2) uint32_t const * ptr, uint32x4x2_t * src, __constrange(0,3) int lane); // VLD2.32 {d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE int16x8x2_t vld2q_lane_s16_ptr(__transfersize(2) int16_t const * ptr, int16x8x2_t * src, __constrange(0,7) int lane); // VLD2.16 {d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE int32x4x2_t vld2q_lane_s32_ptr(__transfersize(2) int32_t const * ptr, int32x4x2_t * src, __constrange(0,3) int lane); // VLD2.32 {d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE float16x8x2_t vld2q_lane_f16_ptr(__transfersize(2) __fp16 const * ptr, float16x8x2_t * src, __constrange(0,7) int lane); // VLD2.16 {d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE float32x4x2_t vld2q_lane_f32_ptr(__transfersize(2) float32_t const * ptr, float32x4x2_t * src, __constrange(0,3) int lane); // VLD2.32 {d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE poly16x8x2_t vld2q_lane_p16_ptr(__transfersize(2) poly16_t const * ptr, poly16x8x2_t * src, __constrange(0,7) int lane); // VLD2.16 {d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE uint8x8x2_t vld2_lane_u8(__transfersize(2) uint8_t const * ptr, uint8x8x2_t src, __constrange(0,7) int lane); //VLD2.8 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE uint16x4x2_t vld2_lane_u16(__transfersize(2) uint16_t const * ptr, uint16x4x2_t src, __constrange(0,3) int lane); // VLD2.16 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE uint32x2x2_t vld2_lane_u32(__transfersize(2) uint32_t const * ptr, uint32x2x2_t src, __constrange(0,1) int lane); // VLD2.32 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE int8x8x2_t vld2_lane_s8(__transfersize(2) int8_t const * ptr, int8x8x2_t src, __constrange(0,7) int lane); //VLD2.8 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE int16x4x2_t vld2_lane_s16(__transfersize(2) int16_t const * ptr, int16x4x2_t src, __constrange(0,3) int lane); //VLD2.16 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE int32x2x2_t vld2_lane_s32(__transfersize(2) int32_t const * ptr, int32x2x2_t src, __constrange(0,1) int lane); //VLD2.32 {d0[0], d1[0]}, [r0]
 //float16x4x2_t vld2_lane_f16_ptr(__transfersize(2) __fp16 const * ptr, float16x4x2_t * src, __constrange(0,3) int lane); // VLD2.16 {d0[0], d1[0]}, [r0]
-float32x2x2_t vld2_lane_f32(__transfersize(2) float32_t const * ptr, float32x2x2_t  src, __constrange(0,1) int lane); // VLD2.32 {d0[0], d1[0]}, [r0]
-poly8x8x2_t vld2_lane_p8(__transfersize(2) poly8_t const * ptr, poly8x8x2_t  src, __constrange(0,7) int lane); //VLD2.8 {d0[0], d1[0]}, [r0]
-poly16x4x2_t vld2_lane_p16(__transfersize(2) poly16_t const * ptr, poly16x4x2_t  src, __constrange(0,3) int lane); // VLD2.16 {d0[0], d1[0]}, [r0]
-uint16x8x3_t vld3q_lane_u16_ptr(__transfersize(3) uint16_t const * ptr, uint16x8x3_t * src, __constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
-uint32x4x3_t vld3q_lane_u32_ptr(__transfersize(3) uint32_t const * ptr, uint32x4x3_t * src, __constrange(0,3) int lane); // VLD3.32 {d0[0], d2[0], d4[0]}, [r0]
-int16x8x3_t vld3q_lane_s16_ptr(__transfersize(3) int16_t const * ptr, int16x8x3_t * src, __constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
-int32x4x3_t vld3q_lane_s32_ptr(__transfersize(3) int32_t const * ptr, int32x4x3_t * src, __constrange(0,3) int lane); // VLD3.32 {d0[0], d2[0], d4[0]}, [r0]
-float16x8x3_t vld3q_lane_f16_ptr(__transfersize(3) __fp16 const * ptr, float16x8x3_t * src, __constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
-float32x4x3_t vld3q_lane_f32_ptr(__transfersize(3) float32_t const * ptr, float32x4x3_t * src, __constrange(0,3) int lane); // VLD3.32 {d0[0], d2[0], d4[0]}, [r0]
-poly16x8x3_t vld3q_lane_p16_ptr(__transfersize(3) poly16_t const * ptr, poly16x8x3_t * src, __constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
-uint8x8x3_t vld3_lane_u8(__transfersize(3) uint8_t const * ptr, uint8x8x3_t src, __constrange(0,7) int lane); //VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
-uint16x4x3_t vld3_lane_u16(__transfersize(3) uint16_t const * ptr, uint16x4x3_t src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
-uint32x2x3_t vld3_lane_u32(__transfersize(3) uint32_t const * ptr, uint32x2x3_t src, __constrange(0,1) int lane); // VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
-int8x8x3_t vld3_lane_s8(__transfersize(3) int8_t const * ptr, int8x8x3_t src, __constrange(0,7) int lane); //VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
-int16x4x3_t vld3_lane_s16(__transfersize(3) int16_t const * ptr, int16x4x3_t src, __constrange(0,3) int lane); //VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
-int32x2x3_t vld3_lane_s32(__transfersize(3) int32_t const * ptr, int32x2x3_t src, __constrange(0,1) int lane); //VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
-float16x4x3_t vld3_lane_f16_ptr(__transfersize(3) __fp16 const * ptr, float16x4x3_t * src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
-float32x2x3_t vld3_lane_f32(__transfersize(3) float32_t const * ptr, float32x2x3_t src, __constrange(0,1) int lane); // VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
-poly8x8x3_t vld3_lane_p8(__transfersize(3) poly8_t const * ptr, poly8x8x3_t src, __constrange(0,7) int lane); //VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
-poly16x4x3_t vld3_lane_p16(__transfersize(3) poly16_t const * ptr, poly16x4x3_t src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
-uint16x8x4_t vld4q_lane_u16_ptr(__transfersize(4) uint16_t const * ptr, uint16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-uint32x4x4_t vld4q_lane_u32_ptr(__transfersize(4) uint32_t const * ptr, uint32x4x4_t * src, __constrange(0,3) int lane); // VLD4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-int16x8x4_t vld4q_lane_s16_ptr(__transfersize(4) int16_t const * ptr, int16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-int32x4x4_t vld4q_lane_s32_ptr(__transfersize(4) int32_t const * ptr, int32x4x4_t * src, __constrange(0,3) int lane); // VLD4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-float16x8x4_t vld4q_lane_f16_ptr(__transfersize(4) __fp16 const * ptr, float16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-float32x4x4_t vld4q_lane_f32_ptr(__transfersize(4) float32_t const * ptr, float32x4x4_t * src, __constrange(0,3) int lane); // VLD4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-poly16x8x4_t vld4q_lane_p16_ptr(__transfersize(4) poly16_t const * ptr, poly16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-uint8x8x4_t vld4_lane_u8(__transfersize(4) uint8_t const * ptr, uint8x8x4_t src, __constrange(0,7) int lane); //VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
-uint16x4x4_t vld4_lane_u16(__transfersize(4) uint16_t const * ptr, uint16x4x4_t src, __constrange(0,3) int lane); // VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
-uint32x2x4_t vld4_lane_u32(__transfersize(4) uint32_t const * ptr, uint32x2x4_t src, __constrange(0,1) int lane); // VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
-int8x8x4_t vld4_lane_s8(__transfersize(4) int8_t const * ptr, int8x8x4_t src, __constrange(0,7) int lane); //VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
-int16x4x4_t vld4_lane_s16(__transfersize(4) int16_t const * ptr, int16x4x4_t src, __constrange(0,3) int lane); //VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
-int32x2x4_t vld4_lane_s32(__transfersize(4) int32_t const * ptr, int32x2x4_t src, __constrange(0,1) int lane); //VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
-float16x4x4_t vld4_lane_f16_ptr(__transfersize(4) __fp16 const * ptr, float16x4x4_t * src, __constrange(0,3) int lane); // VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
-float32x2x4_t vld4_lane_f32(__transfersize(4) float32_t const * ptr, float32x2x4_t src, __constrange(0,1) int lane); // VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
-poly8x8x4_t vld4_lane_p8(__transfersize(4) poly8_t const * ptr, poly8x8x4_t src, __constrange(0,7) int lane); //VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
-poly16x4x4_t vld4_lane_p16(__transfersize(4) poly16_t const * ptr, poly16x4x4_t src, __constrange(0,3) int lane); // VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE float32x2x2_t vld2_lane_f32(__transfersize(2) float32_t const * ptr, float32x2x2_t  src, __constrange(0,1) int lane); // VLD2.32 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE poly8x8x2_t vld2_lane_p8(__transfersize(2) poly8_t const * ptr, poly8x8x2_t  src, __constrange(0,7) int lane); //VLD2.8 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE poly16x4x2_t vld2_lane_p16(__transfersize(2) poly16_t const * ptr, poly16x4x2_t  src, __constrange(0,3) int lane); // VLD2.16 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE uint16x8x3_t vld3q_lane_u16_ptr(__transfersize(3) uint16_t const * ptr, uint16x8x3_t * src, __constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE uint32x4x3_t vld3q_lane_u32_ptr(__transfersize(3) uint32_t const * ptr, uint32x4x3_t * src, __constrange(0,3) int lane); // VLD3.32 {d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE int16x8x3_t vld3q_lane_s16_ptr(__transfersize(3) int16_t const * ptr, int16x8x3_t * src, __constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE int32x4x3_t vld3q_lane_s32_ptr(__transfersize(3) int32_t const * ptr, int32x4x3_t * src, __constrange(0,3) int lane); // VLD3.32 {d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE float16x8x3_t vld3q_lane_f16_ptr(__transfersize(3) __fp16 const * ptr, float16x8x3_t * src, __constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE float32x4x3_t vld3q_lane_f32_ptr(__transfersize(3) float32_t const * ptr, float32x4x3_t * src, __constrange(0,3) int lane); // VLD3.32 {d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE poly16x8x3_t vld3q_lane_p16_ptr(__transfersize(3) poly16_t const * ptr, poly16x8x3_t * src, __constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE uint8x8x3_t vld3_lane_u8(__transfersize(3) uint8_t const * ptr, uint8x8x3_t src, __constrange(0,7) int lane); //VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE uint16x4x3_t vld3_lane_u16(__transfersize(3) uint16_t const * ptr, uint16x4x3_t src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE uint32x2x3_t vld3_lane_u32(__transfersize(3) uint32_t const * ptr, uint32x2x3_t src, __constrange(0,1) int lane); // VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE int8x8x3_t vld3_lane_s8(__transfersize(3) int8_t const * ptr, int8x8x3_t src, __constrange(0,7) int lane); //VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE int16x4x3_t vld3_lane_s16(__transfersize(3) int16_t const * ptr, int16x4x3_t src, __constrange(0,3) int lane); //VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE int32x2x3_t vld3_lane_s32(__transfersize(3) int32_t const * ptr, int32x2x3_t src, __constrange(0,1) int lane); //VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE float16x4x3_t vld3_lane_f16_ptr(__transfersize(3) __fp16 const * ptr, float16x4x3_t * src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE float32x2x3_t vld3_lane_f32(__transfersize(3) float32_t const * ptr, float32x2x3_t src, __constrange(0,1) int lane); // VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE poly8x8x3_t vld3_lane_p8(__transfersize(3) poly8_t const * ptr, poly8x8x3_t src, __constrange(0,7) int lane); //VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE poly16x4x3_t vld3_lane_p16(__transfersize(3) poly16_t const * ptr, poly16x4x3_t src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE uint16x8x4_t vld4q_lane_u16_ptr(__transfersize(4) uint16_t const * ptr, uint16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE uint32x4x4_t vld4q_lane_u32_ptr(__transfersize(4) uint32_t const * ptr, uint32x4x4_t * src, __constrange(0,3) int lane); // VLD4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE int16x8x4_t vld4q_lane_s16_ptr(__transfersize(4) int16_t const * ptr, int16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE int32x4x4_t vld4q_lane_s32_ptr(__transfersize(4) int32_t const * ptr, int32x4x4_t * src, __constrange(0,3) int lane); // VLD4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE float16x8x4_t vld4q_lane_f16_ptr(__transfersize(4) __fp16 const * ptr, float16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE float32x4x4_t vld4q_lane_f32_ptr(__transfersize(4) float32_t const * ptr, float32x4x4_t * src, __constrange(0,3) int lane); // VLD4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE poly16x8x4_t vld4q_lane_p16_ptr(__transfersize(4) poly16_t const * ptr, poly16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE uint8x8x4_t vld4_lane_u8(__transfersize(4) uint8_t const * ptr, uint8x8x4_t src, __constrange(0,7) int lane); //VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE uint16x4x4_t vld4_lane_u16(__transfersize(4) uint16_t const * ptr, uint16x4x4_t src, __constrange(0,3) int lane); // VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE uint32x2x4_t vld4_lane_u32(__transfersize(4) uint32_t const * ptr, uint32x2x4_t src, __constrange(0,1) int lane); // VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE int8x8x4_t vld4_lane_s8(__transfersize(4) int8_t const * ptr, int8x8x4_t src, __constrange(0,7) int lane); //VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE int16x4x4_t vld4_lane_s16(__transfersize(4) int16_t const * ptr, int16x4x4_t src, __constrange(0,3) int lane); //VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE int32x2x4_t vld4_lane_s32(__transfersize(4) int32_t const * ptr, int32x2x4_t src, __constrange(0,1) int lane); //VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE float16x4x4_t vld4_lane_f16_ptr(__transfersize(4) __fp16 const * ptr, float16x4x4_t * src, __constrange(0,3) int lane); // VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE float32x2x4_t vld4_lane_f32(__transfersize(4) float32_t const * ptr, float32x2x4_t src, __constrange(0,1) int lane); // VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE poly8x8x4_t vld4_lane_p8(__transfersize(4) poly8_t const * ptr, poly8x8x4_t src, __constrange(0,7) int lane); //VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE poly16x4x4_t vld4_lane_p16(__transfersize(4) poly16_t const * ptr, poly16x4x4_t src, __constrange(0,3) int lane); // VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 //Store N-element structure to memory
-void vst2q_u8_ptr(__transfersize(32) uint8_t * ptr, uint8x16x2_t * val); // VST2.8 {d0, d2}, [r0]
-void vst2q_u16_ptr(__transfersize(16) uint16_t * ptr, uint16x8x2_t * val); // VST2.16 {d0, d2}, [r0]
-void vst2q_u32_ptr(__transfersize(8) uint32_t * ptr, uint32x4x2_t * val); // VST2.32 {d0, d2}, [r0]
-void vst2q_s8_ptr(__transfersize(32) int8_t * ptr, int8x16x2_t * val); // VST2.8 {d0, d2}, [r0]
-void vst2q_s16_ptr(__transfersize(16) int16_t * ptr, int16x8x2_t * val); // VST2.16 {d0, d2}, [r0]
-void vst2q_s32_ptr(__transfersize(8) int32_t * ptr, int32x4x2_t * val); // VST2.32 {d0, d2}, [r0]
-void vst2q_f16_ptr(__transfersize(16) __fp16 * ptr, float16x8x2_t * val); // VST2.16 {d0, d2}, [r0]
-void vst2q_f32_ptr(__transfersize(8) float32_t * ptr, float32x4x2_t * val); // VST2.32 {d0, d2}, [r0]
-void vst2q_p8_ptr(__transfersize(32) poly8_t * ptr, poly8x16x2_t * val); // VST2.8 {d0, d2}, [r0]
-void vst2q_p16_ptr(__transfersize(16) poly16_t * ptr, poly16x8x2_t * val); // VST2.16 {d0, d2}, [r0]
-void vst2_u8(__transfersize(16) uint8_t * ptr, uint8x8x2_t val); // VST2.8 {d0, d1}, [r0]
-void vst2_u16(__transfersize(8) uint16_t * ptr, uint16x4x2_t val); // VST2.16 {d0, d1}, [r0]
-void vst2_u32(__transfersize(4) uint32_t * ptr, uint32x2x2_t val); // VST2.32 {d0, d1}, [r0]
-void vst2_u64(__transfersize(2) uint64_t * ptr, uint64x1x2_t val); // VST1.64 {d0, d1}, [r0]
-void vst2_s8(__transfersize(16) int8_t * ptr, int8x8x2_t val); // VST2.8 {d0, d1}, [r0]
-void vst2_s16(__transfersize(8) int16_t * ptr, int16x4x2_t val); // VST2.16 {d0, d1}, [r0]
-void vst2_s32(__transfersize(4) int32_t * ptr, int32x2x2_t val); // VST2.32 {d0, d1}, [r0]
-void vst2_s64(__transfersize(2) int64_t * ptr, int64x1x2_t val); // VST1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2q_u8_ptr(__transfersize(32) uint8_t * ptr, uint8x16x2_t * val); // VST2.8 {d0, d2}, [r0]
+_NEON2SSESTORAGE void vst2q_u16_ptr(__transfersize(16) uint16_t * ptr, uint16x8x2_t * val); // VST2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE void vst2q_u32_ptr(__transfersize(8) uint32_t * ptr, uint32x4x2_t * val); // VST2.32 {d0, d2}, [r0]
+_NEON2SSESTORAGE void vst2q_s8_ptr(__transfersize(32) int8_t * ptr, int8x16x2_t * val); // VST2.8 {d0, d2}, [r0]
+_NEON2SSESTORAGE void vst2q_s16_ptr(__transfersize(16) int16_t * ptr, int16x8x2_t * val); // VST2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE void vst2q_s32_ptr(__transfersize(8) int32_t * ptr, int32x4x2_t * val); // VST2.32 {d0, d2}, [r0]
+_NEON2SSESTORAGE void vst2q_f16_ptr(__transfersize(16) __fp16 * ptr, float16x8x2_t * val); // VST2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE void vst2q_f32_ptr(__transfersize(8) float32_t * ptr, float32x4x2_t * val); // VST2.32 {d0, d2}, [r0]
+_NEON2SSESTORAGE void vst2q_p8_ptr(__transfersize(32) poly8_t * ptr, poly8x16x2_t * val); // VST2.8 {d0, d2}, [r0]
+_NEON2SSESTORAGE void vst2q_p16_ptr(__transfersize(16) poly16_t * ptr, poly16x8x2_t * val); // VST2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE void vst2_u8(__transfersize(16) uint8_t * ptr, uint8x8x2_t val); // VST2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_u16(__transfersize(8) uint16_t * ptr, uint16x4x2_t val); // VST2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_u32(__transfersize(4) uint32_t * ptr, uint32x2x2_t val); // VST2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_u64(__transfersize(2) uint64_t * ptr, uint64x1x2_t val); // VST1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_s8(__transfersize(16) int8_t * ptr, int8x8x2_t val); // VST2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_s16(__transfersize(8) int16_t * ptr, int16x4x2_t val); // VST2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_s32(__transfersize(4) int32_t * ptr, int32x2x2_t val); // VST2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_s64(__transfersize(2) int64_t * ptr, int64x1x2_t val); // VST1.64 {d0, d1}, [r0]
 //void vst2_f16_ptr(__transfersize(8) __fp16 * ptr, float16x4x2_t * val); // VST2.16 {d0, d1}, [r0]
-void vst2_f32_ptr(__transfersize(4) float32_t * ptr, float32x2x2_t * val); // VST2.32 {d0, d1}, [r0]
-void vst2_p8(__transfersize(16) poly8_t * ptr, poly8x8x2_t val); // VST2.8 {d0, d1}, [r0]
-void vst2_p16(__transfersize(8) poly16_t * ptr, poly16x4x2_t val); // VST2.16 {d0, d1}, [r0]
-void vst3q_u8_ptr(__transfersize(48) uint8_t * ptr, uint8x16x3_t * val); // VST3.8 {d0, d2, d4}, [r0]
-void vst3q_u16_ptr(__transfersize(24) uint16_t * ptr, uint16x8x3_t * val); // VST3.16 {d0, d2, d4}, [r0]
-void vst3q_u32_ptr(__transfersize(12) uint32_t * ptr, uint32x4x3_t * val); // VST3.32 {d0, d2, d4}, [r0]
-void vst3q_s8_ptr(__transfersize(48) int8_t * ptr, int8x16x3_t * val); // VST3.8 {d0, d2, d4}, [r0]
-void vst3q_s16_ptr(__transfersize(24) int16_t * ptr, int16x8x3_t * val); // VST3.16 {d0, d2, d4}, [r0]
-void vst3q_s32_ptr(__transfersize(12) int32_t * ptr, int32x4x3_t * val); // VST3.32 {d0, d2, d4}, [r0]
-void vst3q_f16_ptr(__transfersize(24) __fp16 * ptr, float16x8x3_t * val); // VST3.16 {d0, d2, d4}, [r0]
-void vst3q_f32_ptr(__transfersize(12) float32_t * ptr, float32x4x3_t * val); // VST3.32 {d0, d2, d4}, [r0]
-void vst3q_p8_ptr(__transfersize(48) poly8_t * ptr, poly8x16x3_t * val); // VST3.8 {d0, d2, d4}, [r0]
-void vst3q_p16_ptr(__transfersize(24) poly16_t * ptr, poly16x8x3_t * val); // VST3.16 {d0, d2, d4}, [r0]
-void vst3_u8(__transfersize(24) uint8_t * ptr, uint8x8x3_t val); // VST3.8 {d0, d1, d2}, [r0]
-void vst3_u16(__transfersize(12) uint16_t * ptr, uint16x4x3_t val); // VST3.16 {d0, d1, d2}, [r0]
-void vst3_u32(__transfersize(6) uint32_t * ptr, uint32x2x3_t val); // VST3.32 {d0, d1, d2}, [r0]
-void vst3_u64(__transfersize(3) uint64_t * ptr, uint64x1x3_t val); // VST1.64 {d0, d1, d2}, [r0]
-void vst3_s8(__transfersize(24) int8_t * ptr, int8x8x3_t val); // VST3.8 {d0, d1, d2}, [r0]
-void vst3_s16(__transfersize(12) int16_t * ptr, int16x4x3_t val); // VST3.16 {d0, d1, d2}, [r0]
-void vst3_s32(__transfersize(6) int32_t * ptr, int32x2x3_t val); // VST3.32 {d0, d1, d2}, [r0]
-void vst3_s64(__transfersize(3) int64_t * ptr, int64x1x3_t val); // VST1.64 {d0, d1, d2}, [r0]
-void vst3_f16_ptr(__transfersize(12) __fp16 * ptr, float16x4x3_t * val); // VST3.16 {d0, d1, d2}, [r0]
-void vst3_f32(__transfersize(6) float32_t * ptr, float32x2x3_t val); // VST3.32 {d0, d1, d2}, [r0]
-void vst3_p8(__transfersize(24) poly8_t * ptr, poly8x8x3_t val); // VST3.8 {d0, d1, d2}, [r0]
-void vst3_p16(__transfersize(12) poly16_t * ptr, poly16x4x3_t val); // VST3.16 {d0, d1, d2}, [r0]
-void vst4q_u8_ptr(__transfersize(64) uint8_t * ptr, uint8x16x4_t * val); // VST4.8 {d0, d2, d4, d6}, [r0]
-void vst4q_u16_ptr(__transfersize(32) uint16_t * ptr, uint16x8x4_t * val); // VST4.16 {d0, d2, d4, d6}, [r0]
-void vst4q_u32_ptr(__transfersize(16) uint32_t * ptr, uint32x4x4_t * val); // VST4.32 {d0, d2, d4, d6}, [r0]
-void vst4q_s8_ptr(__transfersize(64) int8_t * ptr, int8x16x4_t * val); // VST4.8 {d0, d2, d4, d6}, [r0]
-void vst4q_s16_ptr(__transfersize(32) int16_t * ptr, int16x8x4_t * val); // VST4.16 {d0, d2, d4, d6}, [r0]
-void vst4q_s32_ptr(__transfersize(16) int32_t * ptr, int32x4x4_t * val); // VST4.32 {d0, d2, d4, d6}, [r0]
-void vst4q_f16_ptr(__transfersize(32) __fp16 * ptr, float16x8x4_t * val); // VST4.16 {d0, d2, d4, d6}, [r0]
-void vst4q_f32_ptr(__transfersize(16) float32_t * ptr, float32x4x4_t * val); // VST4.32 {d0, d2, d4, d6}, [r0]
-void vst4q_p8_ptr(__transfersize(64) poly8_t * ptr, poly8x16x4_t * val); // VST4.8 {d0, d2, d4, d6}, [r0]
-void vst4q_p16_ptr(__transfersize(32) poly16_t * ptr, poly16x8x4_t * val); // VST4.16 {d0, d2, d4, d6}, [r0]
-void vst4_u8(__transfersize(32) uint8_t * ptr, uint8x8x4_t val); // VST4.8 {d0, d1, d2, d3}, [r0]
-void vst4_u16(__transfersize(16) uint16_t * ptr, uint16x4x4_t val); // VST4.16 {d0, d1, d2, d3}, [r0]
-void vst4_u32(__transfersize(8) uint32_t * ptr, uint32x2x4_t val); // VST4.32 {d0, d1, d2, d3}, [r0]
-void vst4_u64(__transfersize(4) uint64_t * ptr, uint64x1x4_t val); // VST1.64 {d0, d1, d2, d3}, [r0]
-void vst4_s8(__transfersize(32) int8_t * ptr, int8x8x4_t val); // VST4.8 {d0, d1, d2, d3}, [r0]
-void vst4_s16(__transfersize(16) int16_t * ptr, int16x4x4_t val); // VST4.16 {d0, d1, d2, d3}, [r0]
-void vst4_s32(__transfersize(8) int32_t * ptr, int32x2x4_t val); // VST4.32 {d0, d1, d2, d3}, [r0]
-void vst4_s64(__transfersize(4) int64_t * ptr, int64x1x4_t val); // VST1.64 {d0, d1, d2, d3}, [r0]
-void vst4_f16_ptr(__transfersize(16) __fp16 * ptr, float16x4x4_t * val); // VST4.16 {d0, d1, d2, d3}, [r0]
-void vst4_f32(__transfersize(8) float32_t * ptr, float32x2x4_t val); // VST4.32 {d0, d1, d2, d3}, [r0]
-void vst4_p8(__transfersize(32) poly8_t * ptr, poly8x8x4_t val); // VST4.8 {d0, d1, d2, d3}, [r0]
-void vst4_p16(__transfersize(16) poly16_t * ptr, poly16x4x4_t val); // VST4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst2_f32_ptr(__transfersize(4) float32_t * ptr, float32x2x2_t * val); // VST2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_p8(__transfersize(16) poly8_t * ptr, poly8x8x2_t val); // VST2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_p16(__transfersize(8) poly16_t * ptr, poly16x4x2_t val); // VST2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst3q_u8_ptr(__transfersize(48) uint8_t * ptr, uint8x16x3_t * val); // VST3.8 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE void vst3q_u16_ptr(__transfersize(24) uint16_t * ptr, uint16x8x3_t * val); // VST3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE void vst3q_u32_ptr(__transfersize(12) uint32_t * ptr, uint32x4x3_t * val); // VST3.32 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE void vst3q_s8_ptr(__transfersize(48) int8_t * ptr, int8x16x3_t * val); // VST3.8 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE void vst3q_s16_ptr(__transfersize(24) int16_t * ptr, int16x8x3_t * val); // VST3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE void vst3q_s32_ptr(__transfersize(12) int32_t * ptr, int32x4x3_t * val); // VST3.32 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE void vst3q_f16_ptr(__transfersize(24) __fp16 * ptr, float16x8x3_t * val); // VST3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE void vst3q_f32_ptr(__transfersize(12) float32_t * ptr, float32x4x3_t * val); // VST3.32 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE void vst3q_p8_ptr(__transfersize(48) poly8_t * ptr, poly8x16x3_t * val); // VST3.8 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE void vst3q_p16_ptr(__transfersize(24) poly16_t * ptr, poly16x8x3_t * val); // VST3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE void vst3_u8(__transfersize(24) uint8_t * ptr, uint8x8x3_t val); // VST3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_u16(__transfersize(12) uint16_t * ptr, uint16x4x3_t val); // VST3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_u32(__transfersize(6) uint32_t * ptr, uint32x2x3_t val); // VST3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_u64(__transfersize(3) uint64_t * ptr, uint64x1x3_t val); // VST1.64 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_s8(__transfersize(24) int8_t * ptr, int8x8x3_t val); // VST3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_s16(__transfersize(12) int16_t * ptr, int16x4x3_t val); // VST3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_s32(__transfersize(6) int32_t * ptr, int32x2x3_t val); // VST3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_s64(__transfersize(3) int64_t * ptr, int64x1x3_t val); // VST1.64 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_f16_ptr(__transfersize(12) __fp16 * ptr, float16x4x3_t * val); // VST3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_f32(__transfersize(6) float32_t * ptr, float32x2x3_t val); // VST3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_p8(__transfersize(24) poly8_t * ptr, poly8x8x3_t val); // VST3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_p16(__transfersize(12) poly16_t * ptr, poly16x4x3_t val); // VST3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst4q_u8_ptr(__transfersize(64) uint8_t * ptr, uint8x16x4_t * val); // VST4.8 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE void vst4q_u16_ptr(__transfersize(32) uint16_t * ptr, uint16x8x4_t * val); // VST4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE void vst4q_u32_ptr(__transfersize(16) uint32_t * ptr, uint32x4x4_t * val); // VST4.32 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE void vst4q_s8_ptr(__transfersize(64) int8_t * ptr, int8x16x4_t * val); // VST4.8 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE void vst4q_s16_ptr(__transfersize(32) int16_t * ptr, int16x8x4_t * val); // VST4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE void vst4q_s32_ptr(__transfersize(16) int32_t * ptr, int32x4x4_t * val); // VST4.32 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE void vst4q_f16_ptr(__transfersize(32) __fp16 * ptr, float16x8x4_t * val); // VST4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE void vst4q_f32_ptr(__transfersize(16) float32_t * ptr, float32x4x4_t * val); // VST4.32 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE void vst4q_p8_ptr(__transfersize(64) poly8_t * ptr, poly8x16x4_t * val); // VST4.8 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE void vst4q_p16_ptr(__transfersize(32) poly16_t * ptr, poly16x8x4_t * val); // VST4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE void vst4_u8(__transfersize(32) uint8_t * ptr, uint8x8x4_t val); // VST4.8 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_u16(__transfersize(16) uint16_t * ptr, uint16x4x4_t val); // VST4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_u32(__transfersize(8) uint32_t * ptr, uint32x2x4_t val); // VST4.32 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_u64(__transfersize(4) uint64_t * ptr, uint64x1x4_t val); // VST1.64 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_s8(__transfersize(32) int8_t * ptr, int8x8x4_t val); // VST4.8 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_s16(__transfersize(16) int16_t * ptr, int16x4x4_t val); // VST4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_s32(__transfersize(8) int32_t * ptr, int32x2x4_t val); // VST4.32 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_s64(__transfersize(4) int64_t * ptr, int64x1x4_t val); // VST1.64 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_f16_ptr(__transfersize(16) __fp16 * ptr, float16x4x4_t * val); // VST4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_f32(__transfersize(8) float32_t * ptr, float32x2x4_t val); // VST4.32 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_p8(__transfersize(32) poly8_t * ptr, poly8x8x4_t val); // VST4.8 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_p16(__transfersize(16) poly16_t * ptr, poly16x4x4_t val); // VST4.16 {d0, d1, d2, d3}, [r0]
 //Store a single lane of N-element structure to memory
-void vst2q_lane_u16_ptr(__transfersize(2) uint16_t * ptr, uint16x8x2_t * val, __constrange(0,7) int lane); // VST2.16{d0[0], d2[0]}, [r0]
-void vst2q_lane_u32_ptr(__transfersize(2) uint32_t * ptr, uint32x4x2_t * val, __constrange(0,3) int lane); // VST2.32{d0[0], d2[0]}, [r0]
-void vst2q_lane_s16_ptr(__transfersize(2) int16_t * ptr, int16x8x2_t * val, __constrange(0,7) int lane); // VST2.16{d0[0], d2[0]}, [r0]
-void vst2q_lane_s32_ptr(__transfersize(2) int32_t * ptr, int32x4x2_t * val, __constrange(0,3) int lane); // VST2.32{d0[0], d2[0]}, [r0]
-void vst2q_lane_f16_ptr(__transfersize(2) __fp16 * ptr, float16x8x2_t * val, __constrange(0,7) int lane); // VST2.16{d0[0], d2[0]}, [r0]
-void vst2q_lane_f32_ptr(__transfersize(2) float32_t * ptr, float32x4x2_t * val, __constrange(0,3) int lane); //VST2.32 {d0[0], d2[0]}, [r0]
-void vst2q_lane_p16_ptr(__transfersize(2) poly16_t * ptr, poly16x8x2_t * val, __constrange(0,7) int lane); // VST2.16{d0[0], d2[0]}, [r0]
-void vst2_lane_u8(__transfersize(2) uint8_t * ptr, uint8x8x2_t val, __constrange(0,7) int lane); // VST2.8{d0[0], d1[0]}, [r0]
-void vst2_lane_u16(__transfersize(2) uint16_t * ptr, uint16x4x2_t val, __constrange(0,3) int lane); // VST2.16{d0[0], d1[0]}, [r0]
-void vst2_lane_u32(__transfersize(2) uint32_t * ptr, uint32x2x2_t val, __constrange(0,1) int lane); // VST2.32{d0[0], d1[0]}, [r0]
-void vst2_lane_s8(__transfersize(2) int8_t * ptr, int8x8x2_t val, __constrange(0,7) int lane); // VST2.8 {d0[0],d1[0]}, [r0]
-void vst2_lane_s16(__transfersize(2) int16_t * ptr, int16x4x2_t val, __constrange(0,3) int lane); // VST2.16{d0[0], d1[0]}, [r0]
-void vst2_lane_s32(__transfersize(2) int32_t * ptr, int32x2x2_t val, __constrange(0,1) int lane); // VST2.32{d0[0], d1[0]}, [r0]
-void vst2_lane_f16_ptr(__transfersize(2) __fp16 * ptr, float16x4x2_t * val, __constrange(0,3) int lane); // VST2.16{d0[0], d1[0]}, [r0]
-void vst2_lane_f32(__transfersize(2) float32_t * ptr, float32x2x2_t val, __constrange(0,1) int lane); // VST2.32{d0[0], d1[0]}, [r0]
-void vst2_lane_p8(__transfersize(2) poly8_t * ptr, poly8x8x2_t val, __constrange(0,7) int lane); // VST2.8{d0[0], d1[0]}, [r0]
-void vst2_lane_p16(__transfersize(2) poly16_t * ptr, poly16x4x2_t val, __constrange(0,3) int lane); // VST2.16{d0[0], d1[0]}, [r0]
-void vst3q_lane_u16_ptr(__transfersize(3) uint16_t * ptr, uint16x8x3_t * val, __constrange(0,7) int lane); // VST3.16{d0[0], d2[0], d4[0]}, [r0]
-void vst3q_lane_u32_ptr(__transfersize(3) uint32_t * ptr, uint32x4x3_t * val, __constrange(0,3) int lane); // VST3.32{d0[0], d2[0], d4[0]}, [r0]
-void vst3q_lane_s16_ptr(__transfersize(3) int16_t * ptr, int16x8x3_t * val, __constrange(0,7) int lane); // VST3.16{d0[0], d2[0], d4[0]}, [r0]
-void vst3q_lane_s32_ptr(__transfersize(3) int32_t * ptr, int32x4x3_t * val, __constrange(0,3) int lane); // VST3.32{d0[0], d2[0], d4[0]}, [r0]
-void vst3q_lane_f16_ptr(__transfersize(3) __fp16 * ptr, float16x8x3_t * val, __constrange(0,7) int lane); // VST3.16{d0[0], d2[0], d4[0]}, [r0]
-void vst3q_lane_f32_ptr(__transfersize(3) float32_t * ptr, float32x4x3_t * val, __constrange(0,3) int lane); //VST3.32 {d0[0], d2[0], d4[0]}, [r0]
-void vst3q_lane_p16_ptr(__transfersize(3) poly16_t * ptr, poly16x8x3_t * val, __constrange(0,7) int lane); // VST3.16{d0[0], d2[0], d4[0]}, [r0]
-void vst3_lane_u8(__transfersize(3) uint8_t * ptr, uint8x8x3_t val, __constrange(0,7) int lane); // VST3.8{d0[0], d1[0], d2[0]}, [r0]
-void vst3_lane_u16(__transfersize(3) uint16_t * ptr, uint16x4x3_t val, __constrange(0,3) int lane); // VST3.16{d0[0], d1[0], d2[0]}, [r0]
-void vst3_lane_u32(__transfersize(3) uint32_t * ptr, uint32x2x3_t val, __constrange(0,1) int lane); // VST3.32{d0[0], d1[0], d2[0]}, [r0]
-void vst3_lane_s8(__transfersize(3) int8_t * ptr, int8x8x3_t val, __constrange(0,7) int lane); // VST3.8 {d0[0],d1[0], d2[0]}, [r0]
-void vst3_lane_s16(__transfersize(3) int16_t * ptr, int16x4x3_t val, __constrange(0,3) int lane); // VST3.16{d0[0], d1[0], d2[0]}, [r0]
-void vst3_lane_s32(__transfersize(3) int32_t * ptr, int32x2x3_t val, __constrange(0,1) int lane); // VST3.32{d0[0], d1[0], d2[0]}, [r0]
-void vst3_lane_f16_ptr(__transfersize(3) __fp16 * ptr, float16x4x3_t * val, __constrange(0,3) int lane); // VST3.16{d0[0], d1[0], d2[0]}, [r0]
-void vst3_lane_f32(__transfersize(3) float32_t * ptr, float32x2x3_t val, __constrange(0,1) int lane); // VST3.32{d0[0], d1[0], d2[0]}, [r0]
-void vst3_lane_p8(__transfersize(3) poly8_t * ptr, poly8x8x3_t val, __constrange(0,7) int lane); // VST3.8{d0[0], d1[0], d2[0]}, [r0]
-void vst3_lane_p16(__transfersize(3) poly16_t * ptr, poly16x4x3_t val, __constrange(0,3) int lane); // VST3.16{d0[0], d1[0], d2[0]}, [r0]
-void vst4q_lane_u16_ptr(__transfersize(4) uint16_t * ptr, uint16x8x4_t * val, __constrange(0,7) int lane); // VST4.16{d0[0], d2[0], d4[0], d6[0]}, [r0]
-void vst4q_lane_u32_ptr(__transfersize(4) uint32_t * ptr, uint32x4x4_t * val, __constrange(0,3) int lane); // VST4.32{d0[0], d2[0], d4[0], d6[0]}, [r0]
-void vst4q_lane_s16_ptr(__transfersize(4) int16_t * ptr, int16x8x4_t * val, __constrange(0,7) int lane); // VST4.16{d0[0], d2[0], d4[0], d6[0]}, [r0]
-void vst4q_lane_s32_ptr(__transfersize(4) int32_t * ptr, int32x4x4_t * val, __constrange(0,3) int lane); // VST4.32{d0[0], d2[0], d4[0], d6[0]}, [r0]
-void vst4q_lane_f16_ptr(__transfersize(4) __fp16 * ptr, float16x8x4_t * val, __constrange(0,7) int lane); // VST4.16{d0[0], d2[0], d4[0], d6[0]}, [r0]
-void vst4q_lane_f32_ptr(__transfersize(4) float32_t * ptr, float32x4x4_t * val, __constrange(0,3) int lane); //VST4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-void vst4q_lane_p16_ptr(__transfersize(4) poly16_t * ptr, poly16x8x4_t * val, __constrange(0,7) int lane); // VST4.16{d0[0], d2[0], d4[0], d6[0]}, [r0]
-void vst4_lane_u8(__transfersize(4) uint8_t * ptr, uint8x8x4_t val, __constrange(0,7) int lane); // VST4.8{d0[0], d1[0], d2[0], d3[0]}, [r0]
-void vst4_lane_u16(__transfersize(4) uint16_t * ptr, uint16x4x4_t val, __constrange(0,3) int lane); // VST4.16{d0[0], d1[0], d2[0], d3[0]}, [r0]
-void vst4_lane_u32(__transfersize(4) uint32_t * ptr, uint32x2x4_t val, __constrange(0,1) int lane); // VST4.32{d0[0], d1[0], d2[0], d3[0]}, [r0]
-void vst4_lane_s8(__transfersize(4) int8_t * ptr, int8x8x4_t val, __constrange(0,7) int lane); // VST4.8 {d0[0],d1[0], d2[0], d3[0]}, [r0]
-void vst4_lane_s16(__transfersize(4) int16_t * ptr, int16x4x4_t val, __constrange(0,3) int lane); // VST4.16{d0[0], d1[0], d2[0], d3[0]}, [r0]
-void vst4_lane_s32(__transfersize(4) int32_t * ptr, int32x2x4_t val, __constrange(0,1) int lane); // VST4.32{d0[0], d1[0], d2[0], d3[0]}, [r0]
-void vst4_lane_f16_ptr(__transfersize(4) __fp16 * ptr, float16x4x4_t * val, __constrange(0,3) int lane); // VST4.16{d0[0], d1[0], d2[0], d3[0]}, [r0]
-void vst4_lane_f32(__transfersize(4) float32_t * ptr, float32x2x4_t val, __constrange(0,1) int lane); // VST4.32{d0[0], d1[0], d2[0], d3[0]}, [r0]
-void vst4_lane_p8(__transfersize(4) poly8_t * ptr, poly8x8x4_t val, __constrange(0,7) int lane); // VST4.8{d0[0], d1[0], d2[0], d3[0]}, [r0]
-void vst4_lane_p16(__transfersize(4) poly16_t * ptr, poly16x4x4_t val, __constrange(0,3) int lane); // VST4.16{d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst2q_lane_u16_ptr(__transfersize(2) uint16_t * ptr, uint16x8x2_t * val, __constrange(0,7) int lane); // VST2.16{d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst2q_lane_u32_ptr(__transfersize(2) uint32_t * ptr, uint32x4x2_t * val, __constrange(0,3) int lane); // VST2.32{d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst2q_lane_s16_ptr(__transfersize(2) int16_t * ptr, int16x8x2_t * val, __constrange(0,7) int lane); // VST2.16{d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst2q_lane_s32_ptr(__transfersize(2) int32_t * ptr, int32x4x2_t * val, __constrange(0,3) int lane); // VST2.32{d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst2q_lane_f16_ptr(__transfersize(2) __fp16 * ptr, float16x8x2_t * val, __constrange(0,7) int lane); // VST2.16{d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst2q_lane_f32_ptr(__transfersize(2) float32_t * ptr, float32x4x2_t * val, __constrange(0,3) int lane); //VST2.32 {d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst2q_lane_p16_ptr(__transfersize(2) poly16_t * ptr, poly16x8x2_t * val, __constrange(0,7) int lane); // VST2.16{d0[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_u8(__transfersize(2) uint8_t * ptr, uint8x8x2_t val, __constrange(0,7) int lane); // VST2.8{d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_u16(__transfersize(2) uint16_t * ptr, uint16x4x2_t val, __constrange(0,3) int lane); // VST2.16{d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_u32(__transfersize(2) uint32_t * ptr, uint32x2x2_t val, __constrange(0,1) int lane); // VST2.32{d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_s8(__transfersize(2) int8_t * ptr, int8x8x2_t val, __constrange(0,7) int lane); // VST2.8 {d0[0],d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_s16(__transfersize(2) int16_t * ptr, int16x4x2_t val, __constrange(0,3) int lane); // VST2.16{d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_s32(__transfersize(2) int32_t * ptr, int32x2x2_t val, __constrange(0,1) int lane); // VST2.32{d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_f16_ptr(__transfersize(2) __fp16 * ptr, float16x4x2_t * val, __constrange(0,3) int lane); // VST2.16{d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_f32(__transfersize(2) float32_t * ptr, float32x2x2_t val, __constrange(0,1) int lane); // VST2.32{d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_p8(__transfersize(2) poly8_t * ptr, poly8x8x2_t val, __constrange(0,7) int lane); // VST2.8{d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_p16(__transfersize(2) poly16_t * ptr, poly16x4x2_t val, __constrange(0,3) int lane); // VST2.16{d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst3q_lane_u16_ptr(__transfersize(3) uint16_t * ptr, uint16x8x3_t * val, __constrange(0,7) int lane); // VST3.16{d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE void vst3q_lane_u32_ptr(__transfersize(3) uint32_t * ptr, uint32x4x3_t * val, __constrange(0,3) int lane); // VST3.32{d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE void vst3q_lane_s16_ptr(__transfersize(3) int16_t * ptr, int16x8x3_t * val, __constrange(0,7) int lane); // VST3.16{d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE void vst3q_lane_s32_ptr(__transfersize(3) int32_t * ptr, int32x4x3_t * val, __constrange(0,3) int lane); // VST3.32{d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE void vst3q_lane_f16_ptr(__transfersize(3) __fp16 * ptr, float16x8x3_t * val, __constrange(0,7) int lane); // VST3.16{d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE void vst3q_lane_f32_ptr(__transfersize(3) float32_t * ptr, float32x4x3_t * val, __constrange(0,3) int lane); //VST3.32 {d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE void vst3q_lane_p16_ptr(__transfersize(3) poly16_t * ptr, poly16x8x3_t * val, __constrange(0,7) int lane); // VST3.16{d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_u8(__transfersize(3) uint8_t * ptr, uint8x8x3_t val, __constrange(0,7) int lane); // VST3.8{d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_u16(__transfersize(3) uint16_t * ptr, uint16x4x3_t val, __constrange(0,3) int lane); // VST3.16{d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_u32(__transfersize(3) uint32_t * ptr, uint32x2x3_t val, __constrange(0,1) int lane); // VST3.32{d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_s8(__transfersize(3) int8_t * ptr, int8x8x3_t val, __constrange(0,7) int lane); // VST3.8 {d0[0],d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_s16(__transfersize(3) int16_t * ptr, int16x4x3_t val, __constrange(0,3) int lane); // VST3.16{d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_s32(__transfersize(3) int32_t * ptr, int32x2x3_t val, __constrange(0,1) int lane); // VST3.32{d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_f16_ptr(__transfersize(3) __fp16 * ptr, float16x4x3_t * val, __constrange(0,3) int lane); // VST3.16{d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_f32(__transfersize(3) float32_t * ptr, float32x2x3_t val, __constrange(0,1) int lane); // VST3.32{d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_p8(__transfersize(3) poly8_t * ptr, poly8x8x3_t val, __constrange(0,7) int lane); // VST3.8{d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_p16(__transfersize(3) poly16_t * ptr, poly16x4x3_t val, __constrange(0,3) int lane); // VST3.16{d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst4q_lane_u16_ptr(__transfersize(4) uint16_t * ptr, uint16x8x4_t * val, __constrange(0,7) int lane); // VST4.16{d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE void vst4q_lane_u32_ptr(__transfersize(4) uint32_t * ptr, uint32x4x4_t * val, __constrange(0,3) int lane); // VST4.32{d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE void vst4q_lane_s16_ptr(__transfersize(4) int16_t * ptr, int16x8x4_t * val, __constrange(0,7) int lane); // VST4.16{d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE void vst4q_lane_s32_ptr(__transfersize(4) int32_t * ptr, int32x4x4_t * val, __constrange(0,3) int lane); // VST4.32{d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE void vst4q_lane_f16_ptr(__transfersize(4) __fp16 * ptr, float16x8x4_t * val, __constrange(0,7) int lane); // VST4.16{d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE void vst4q_lane_f32_ptr(__transfersize(4) float32_t * ptr, float32x4x4_t * val, __constrange(0,3) int lane); //VST4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE void vst4q_lane_p16_ptr(__transfersize(4) poly16_t * ptr, poly16x8x4_t * val, __constrange(0,7) int lane); // VST4.16{d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_u8(__transfersize(4) uint8_t * ptr, uint8x8x4_t val, __constrange(0,7) int lane); // VST4.8{d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_u16(__transfersize(4) uint16_t * ptr, uint16x4x4_t val, __constrange(0,3) int lane); // VST4.16{d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_u32(__transfersize(4) uint32_t * ptr, uint32x2x4_t val, __constrange(0,1) int lane); // VST4.32{d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_s8(__transfersize(4) int8_t * ptr, int8x8x4_t val, __constrange(0,7) int lane); // VST4.8 {d0[0],d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_s16(__transfersize(4) int16_t * ptr, int16x4x4_t val, __constrange(0,3) int lane); // VST4.16{d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_s32(__transfersize(4) int32_t * ptr, int32x2x4_t val, __constrange(0,1) int lane); // VST4.32{d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_f16_ptr(__transfersize(4) __fp16 * ptr, float16x4x4_t * val, __constrange(0,3) int lane); // VST4.16{d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_f32(__transfersize(4) float32_t * ptr, float32x2x4_t val, __constrange(0,1) int lane); // VST4.32{d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_p8(__transfersize(4) poly8_t * ptr, poly8x8x4_t val, __constrange(0,7) int lane); // VST4.8{d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_p16(__transfersize(4) poly16_t * ptr, poly16x4x4_t val, __constrange(0,3) int lane); // VST4.16{d0[0], d1[0], d2[0], d3[0]}, [r0]
 //Extract lanes from a vector and put into a register. These intrinsics extract a single lane (element) from a vector.
-uint8_t vget_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VMOV.U8 r0, d0[0]
-uint16_t vget_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VMOV.U16 r0, d0[0]
-uint32_t vget_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
-int8_t vget_lane_s8(int8x8_t vec, __constrange(0,7) int lane); // VMOV.S8 r0, d0[0]
-int16_t vget_lane_s16(int16x4_t vec, __constrange(0,3) int lane); // VMOV.S16 r0, d0[0]
-int32_t vget_lane_s32(int32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
-poly8_t vget_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VMOV.U8 r0, d0[0]
-poly16_t vget_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VMOV.U16 r0, d0[0]
-float32_t vget_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
-uint8_t vgetq_lane_u8(uint8x16_t vec, __constrange(0,15) int lane); // VMOV.U8 r0, d0[0]
-uint16_t vgetq_lane_u16(uint16x8_t vec, __constrange(0,7) int lane); // VMOV.U16 r0, d0[0]
-uint32_t vgetq_lane_u32(uint32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
-int8_t vgetq_lane_s8(int8x16_t vec, __constrange(0,15) int lane); // VMOV.S8 r0, d0[0]
-int16_t vgetq_lane_s16(int16x8_t vec, __constrange(0,7) int lane); // VMOV.S16 r0, d0[0]
-int32_t vgetq_lane_s32(int32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
-poly8_t vgetq_lane_p8(poly8x16_t vec, __constrange(0,15) int lane); // VMOV.U8 r0, d0[0]
-poly16_t vgetq_lane_p16(poly16x8_t vec, __constrange(0,7) int lane); // VMOV.U16 r0, d0[0]
-float32_t vgetq_lane_f32(float32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
-int64_t vget_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV r0,r0,d0
-uint64_t vget_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV r0,r0,d0
-int64_t vgetq_lane_s64(int64x2_t vec, __constrange(0,1) int lane); // VMOV r0,r0,d0
-uint64_t vgetq_lane_u64(uint64x2_t vec, __constrange(0,1) int lane); // VMOV r0,r0,d0
+_NEON2SSESTORAGE uint8_t vget_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VMOV.U8 r0, d0[0]
+_NEON2SSESTORAGE uint16_t vget_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VMOV.U16 r0, d0[0]
+_NEON2SSESTORAGE uint32_t vget_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE int8_t vget_lane_s8(int8x8_t vec, __constrange(0,7) int lane); // VMOV.S8 r0, d0[0]
+_NEON2SSESTORAGE int16_t vget_lane_s16(int16x4_t vec, __constrange(0,3) int lane); // VMOV.S16 r0, d0[0]
+_NEON2SSESTORAGE int32_t vget_lane_s32(int32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE poly8_t vget_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VMOV.U8 r0, d0[0]
+_NEON2SSESTORAGE poly16_t vget_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VMOV.U16 r0, d0[0]
+_NEON2SSESTORAGE float32_t vget_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE uint8_t vgetq_lane_u8(uint8x16_t vec, __constrange(0,15) int lane); // VMOV.U8 r0, d0[0]
+_NEON2SSESTORAGE uint16_t vgetq_lane_u16(uint16x8_t vec, __constrange(0,7) int lane); // VMOV.U16 r0, d0[0]
+_NEON2SSESTORAGE uint32_t vgetq_lane_u32(uint32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE int8_t vgetq_lane_s8(int8x16_t vec, __constrange(0,15) int lane); // VMOV.S8 r0, d0[0]
+_NEON2SSESTORAGE int16_t vgetq_lane_s16(int16x8_t vec, __constrange(0,7) int lane); // VMOV.S16 r0, d0[0]
+_NEON2SSESTORAGE int32_t vgetq_lane_s32(int32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE poly8_t vgetq_lane_p8(poly8x16_t vec, __constrange(0,15) int lane); // VMOV.U8 r0, d0[0]
+_NEON2SSESTORAGE poly16_t vgetq_lane_p16(poly16x8_t vec, __constrange(0,7) int lane); // VMOV.U16 r0, d0[0]
+_NEON2SSESTORAGE float32_t vgetq_lane_f32(float32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE int64_t vget_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV r0,r0,d0
+_NEON2SSESTORAGE uint64_t vget_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV r0,r0,d0
+_NEON2SSESTORAGE int64_t vgetq_lane_s64(int64x2_t vec, __constrange(0,1) int lane); // VMOV r0,r0,d0
+_NEON2SSESTORAGE uint64_t vgetq_lane_u64(uint64x2_t vec, __constrange(0,1) int lane); // VMOV r0,r0,d0
 //Load a single lane of a vector from a literal. These intrinsics set a single lane (element) within a vector.
-uint8x8_t vset_lane_u8(uint8_t value, uint8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
-uint16x4_t vset_lane_u16(uint16_t value, uint16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
-uint32x2_t vset_lane_u32(uint32_t value, uint32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
-int8x8_t vset_lane_s8(int8_t value, int8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
-int16x4_t vset_lane_s16(int16_t value, int16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
-int32x2_t vset_lane_s32(int32_t value, int32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
-poly8x8_t vset_lane_p8(poly8_t value, poly8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
-poly16x4_t vset_lane_p16(poly16_t value, poly16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
-float32x2_t vset_lane_f32(float32_t value, float32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
-uint8x16_t vsetq_lane_u8(uint8_t value, uint8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
-uint16x8_t vsetq_lane_u16(uint16_t value, uint16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
-uint32x4_t vsetq_lane_u32(uint32_t value, uint32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
-int8x16_t vsetq_lane_s8(int8_t value, int8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
-int16x8_t vsetq_lane_s16(int16_t value, int16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
-int32x4_t vsetq_lane_s32(int32_t value, int32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
-poly8x16_t vsetq_lane_p8(poly8_t value, poly8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
-poly16x8_t vsetq_lane_p16(poly16_t value, poly16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
-float32x4_t vsetq_lane_f32(float32_t value, float32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
-int64x1_t vset_lane_s64(int64_t value, int64x1_t vec, __constrange(0,0) int lane); // VMOV d0,r0,r0
-uint64x1_t vset_lane_u64(uint64_t value, uint64x1_t vec, __constrange(0,0) int lane); // VMOV d0,r0,r0
-int64x2_t vsetq_lane_s64(int64_t value, int64x2_t vec, __constrange(0,1) int lane); // VMOV d0,r0,r0
-uint64x2_t vsetq_lane_u64(uint64_t value, uint64x2_t vec, __constrange(0,1) int lane); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint8x8_t vset_lane_u8(uint8_t value, uint8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE uint16x4_t vset_lane_u16(uint16_t value, uint16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE uint32x2_t vset_lane_u32(uint32_t value, uint32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE int8x8_t vset_lane_s8(int8_t value, int8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE int16x4_t vset_lane_s16(int16_t value, int16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE int32x2_t vset_lane_s32(int32_t value, int32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE poly8x8_t vset_lane_p8(poly8_t value, poly8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE poly16x4_t vset_lane_p16(poly16_t value, poly16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE float32x2_t vset_lane_f32(float32_t value, float32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE uint8x16_t vsetq_lane_u8(uint8_t value, uint8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE uint16x8_t vsetq_lane_u16(uint16_t value, uint16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE uint32x4_t vsetq_lane_u32(uint32_t value, uint32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE int8x16_t vsetq_lane_s8(int8_t value, int8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE int16x8_t vsetq_lane_s16(int16_t value, int16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE int32x4_t vsetq_lane_s32(int32_t value, int32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE poly8x16_t vsetq_lane_p8(poly8_t value, poly8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE poly16x8_t vsetq_lane_p16(poly16_t value, poly16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE float32x4_t vsetq_lane_f32(float32_t value, float32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE int64x1_t vset_lane_s64(int64_t value, int64x1_t vec, __constrange(0,0) int lane); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x1_t vset_lane_u64(uint64_t value, uint64x1_t vec, __constrange(0,0) int lane); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int64x2_t vsetq_lane_s64(int64_t value, int64x2_t vec, __constrange(0,1) int lane); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x2_t vsetq_lane_u64(uint64_t value, uint64x2_t vec, __constrange(0,1) int lane); // VMOV d0,r0,r0
 //Initialize a vector from a literal bit pattern.
-int8x8_t vcreate_s8(uint64_t a); // VMOV d0,r0,r0
-int16x4_t vcreate_s16(uint64_t a); // VMOV d0,r0,r0
-int32x2_t vcreate_s32(uint64_t a); // VMOV d0,r0,r0
-float16x4_t vcreate_f16(uint64_t a); // VMOV d0,r0,r0
-float32x2_t vcreate_f32(uint64_t a); // VMOV d0,r0,r0
-uint8x8_t vcreate_u8(uint64_t a); // VMOV d0,r0,r0
-uint16x4_t vcreate_u16(uint64_t a); // VMOV d0,r0,r0
-uint32x2_t vcreate_u32(uint64_t a); // VMOV d0,r0,r0
-uint64x1_t vcreate_u64(uint64_t a); // VMOV d0,r0,r0
-poly8x8_t vcreate_p8(uint64_t a); // VMOV d0,r0,r0
-poly16x4_t vcreate_p16(uint64_t a); // VMOV d0,r0,r0
-int64x1_t vcreate_s64(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int8x8_t vcreate_s8(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int16x4_t vcreate_s16(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int32x2_t vcreate_s32(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE float16x4_t vcreate_f16(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE float32x2_t vcreate_f32(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint8x8_t vcreate_u8(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint16x4_t vcreate_u16(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint32x2_t vcreate_u32(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x1_t vcreate_u64(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE poly8x8_t vcreate_p8(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE poly16x4_t vcreate_p16(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int64x1_t vcreate_s64(uint64_t a); // VMOV d0,r0,r0
 //Set all lanes to same value
 //Load all lanes of vector to the same literal value
-uint8x8_t vdup_n_u8(uint8_t value); // VDUP.8 d0,r0
-uint16x4_t vdup_n_u16(uint16_t value); // VDUP.16 d0,r0
-uint32x2_t vdup_n_u32(uint32_t value); // VDUP.32 d0,r0
-int8x8_t vdup_n_s8(int8_t value); // VDUP.8 d0,r0
-int16x4_t vdup_n_s16(int16_t value); // VDUP.16 d0,r0
-int32x2_t vdup_n_s32(int32_t value); // VDUP.32 d0,r0
-poly8x8_t vdup_n_p8(poly8_t value); // VDUP.8 d0,r0
-poly16x4_t vdup_n_p16(poly16_t value); // VDUP.16 d0,r0
-float32x2_t vdup_n_f32(float32_t value); // VDUP.32 d0,r0
-uint8x16_t vdupq_n_u8(uint8_t value); // VDUP.8 q0,r0
-uint16x8_t vdupq_n_u16(uint16_t value); // VDUP.16 q0,r0
-uint32x4_t vdupq_n_u32(uint32_t value); // VDUP.32 q0,r0
-int8x16_t vdupq_n_s8(int8_t value); // VDUP.8 q0,r0
-int16x8_t vdupq_n_s16(int16_t value); // VDUP.16 q0,r0
-int32x4_t vdupq_n_s32(int32_t value); // VDUP.32 q0,r0
-poly8x16_t vdupq_n_p8(poly8_t value); // VDUP.8 q0,r0
-poly16x8_t vdupq_n_p16(poly16_t value); // VDUP.16 q0,r0
-float32x4_t vdupq_n_f32(float32_t value); // VDUP.32 q0,r0
-int64x1_t vdup_n_s64(int64_t value); // VMOV d0,r0,r0
-uint64x1_t vdup_n_u64(uint64_t value); // VMOV d0,r0,r0
-int64x2_t vdupq_n_s64(int64_t value); // VMOV d0,r0,r0
-uint64x2_t vdupq_n_u64(uint64_t value); // VMOV d0,r0,r0
-uint8x8_t vmov_n_u8(uint8_t value); // VDUP.8 d0,r0
-uint16x4_t vmov_n_u16(uint16_t value); // VDUP.16 d0,r0
-uint32x2_t vmov_n_u32(uint32_t value); // VDUP.32 d0,r0
-int8x8_t vmov_n_s8(int8_t value); // VDUP.8 d0,r0
-int16x4_t vmov_n_s16(int16_t value); // VDUP.16 d0,r0
-int32x2_t vmov_n_s32(int32_t value); // VDUP.32 d0,r0
-poly8x8_t vmov_n_p8(poly8_t value); // VDUP.8 d0,r0
-poly16x4_t vmov_n_p16(poly16_t value); // VDUP.16 d0,r0
-float32x2_t vmov_n_f32(float32_t value); // VDUP.32 d0,r0
-uint8x16_t vmovq_n_u8(uint8_t value); // VDUP.8 q0,r0
-uint16x8_t vmovq_n_u16(uint16_t value); // VDUP.16 q0,r0
-uint32x4_t vmovq_n_u32(uint32_t value); // VDUP.32 q0,r0
-int8x16_t vmovq_n_s8(int8_t value); // VDUP.8 q0,r0
-int16x8_t vmovq_n_s16(int16_t value); // VDUP.16 q0,r0
-int32x4_t vmovq_n_s32(int32_t value); // VDUP.32 q0,r0
-poly8x16_t vmovq_n_p8(poly8_t value); // VDUP.8 q0,r0
-poly16x8_t vmovq_n_p16(poly16_t value); // VDUP.16 q0,r0
-float32x4_t vmovq_n_f32(float32_t value); // VDUP.32 q0,r0
-int64x1_t vmov_n_s64(int64_t value); // VMOV d0,r0,r0
-uint64x1_t vmov_n_u64(uint64_t value); // VMOV d0,r0,r0
-int64x2_t vmovq_n_s64(int64_t value); // VMOV d0,r0,r0
-uint64x2_t vmovq_n_u64(uint64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint8x8_t vdup_n_u8(uint8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE uint16x4_t vdup_n_u16(uint16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE uint32x2_t vdup_n_u32(uint32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE int8x8_t vdup_n_s8(int8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE int16x4_t vdup_n_s16(int16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE int32x2_t vdup_n_s32(int32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE poly8x8_t vdup_n_p8(poly8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE poly16x4_t vdup_n_p16(poly16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE float32x2_t vdup_n_f32(float32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE uint8x16_t vdupq_n_u8(uint8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE uint16x8_t vdupq_n_u16(uint16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE uint32x4_t vdupq_n_u32(uint32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE int8x16_t vdupq_n_s8(int8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE int16x8_t vdupq_n_s16(int16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE int32x4_t vdupq_n_s32(int32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE poly8x16_t vdupq_n_p8(poly8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE poly16x8_t vdupq_n_p16(poly16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE float32x4_t vdupq_n_f32(float32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE int64x1_t vdup_n_s64(int64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x1_t vdup_n_u64(uint64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int64x2_t vdupq_n_s64(int64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x2_t vdupq_n_u64(uint64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint8x8_t vmov_n_u8(uint8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE uint16x4_t vmov_n_u16(uint16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE uint32x2_t vmov_n_u32(uint32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE int8x8_t vmov_n_s8(int8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE int16x4_t vmov_n_s16(int16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE int32x2_t vmov_n_s32(int32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE poly8x8_t vmov_n_p8(poly8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE poly16x4_t vmov_n_p16(poly16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE float32x2_t vmov_n_f32(float32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE uint8x16_t vmovq_n_u8(uint8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE uint16x8_t vmovq_n_u16(uint16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE uint32x4_t vmovq_n_u32(uint32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE int8x16_t vmovq_n_s8(int8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE int16x8_t vmovq_n_s16(int16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE int32x4_t vmovq_n_s32(int32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE poly8x16_t vmovq_n_p8(poly8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE poly16x8_t vmovq_n_p16(poly16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE float32x4_t vmovq_n_f32(float32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE int64x1_t vmov_n_s64(int64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x1_t vmov_n_u64(uint64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int64x2_t vmovq_n_s64(int64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x2_t vmovq_n_u64(uint64_t value); // VMOV d0,r0,r0
 //Load all lanes of the vector to the value of a lane of a vector
-uint8x8_t vdup_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
-uint16x4_t vdup_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
-uint32x2_t vdup_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
-int8x8_t vdup_lane_s8(int8x8_t vec, __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
-int16x4_t vdup_lane_s16(int16x4_t vec, __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
-int32x2_t vdup_lane_s32(int32x2_t vec, __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
-poly8x8_t vdup_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
-poly16x4_t vdup_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
-float32x2_t vdup_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
-uint8x16_t vdupq_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
-uint16x8_t vdupq_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
-uint32x4_t vdupq_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
-int8x16_t vdupq_lane_s8(int8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
-int16x8_t vdupq_lane_s16(int16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
-int32x4_t vdupq_lane_s32(int32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
-poly8x16_t vdupq_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
-poly16x8_t vdupq_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
-float32x4_t vdupq_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
-int64x1_t vdup_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV d0,d0
-uint64x1_t vdup_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV d0,d0
-int64x2_t vdupq_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV q0,q0
-uint64x2_t vdupq_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV q0,q0
+_NEON2SSESTORAGE uint8x8_t vdup_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
+_NEON2SSESTORAGE uint16x4_t vdup_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
+_NEON2SSESTORAGE uint32x2_t vdup_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
+_NEON2SSESTORAGE int8x8_t vdup_lane_s8(int8x8_t vec, __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
+_NEON2SSESTORAGE int16x4_t vdup_lane_s16(int16x4_t vec, __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vdup_lane_s32(int32x2_t vec, __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
+_NEON2SSESTORAGE poly8x8_t vdup_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
+_NEON2SSESTORAGE poly16x4_t vdup_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
+_NEON2SSESTORAGE float32x2_t vdup_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
+_NEON2SSESTORAGE uint8x16_t vdupq_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
+_NEON2SSESTORAGE uint16x8_t vdupq_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
+_NEON2SSESTORAGE uint32x4_t vdupq_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
+_NEON2SSESTORAGE int8x16_t vdupq_lane_s8(int8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
+_NEON2SSESTORAGE int16x8_t vdupq_lane_s16(int16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vdupq_lane_s32(int32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
+_NEON2SSESTORAGE poly8x16_t vdupq_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
+_NEON2SSESTORAGE poly16x8_t vdupq_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
+_NEON2SSESTORAGE float32x4_t vdupq_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
+_NEON2SSESTORAGE int64x1_t vdup_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV d0,d0
+_NEON2SSESTORAGE uint64x1_t vdup_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV d0,d0
+_NEON2SSESTORAGE int64x2_t vdupq_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV q0,q0
+_NEON2SSESTORAGE uint64x2_t vdupq_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV q0,q0
 //Combining vectors. These intrinsics join two 64 bit vectors into a single 128bit vector.
-int8x16_t vcombine_s8(int8x8_t low, int8x8_t high); // VMOV d0,d0
-int16x8_t vcombine_s16(int16x4_t low, int16x4_t high); // VMOV d0,d0
-int32x4_t vcombine_s32(int32x2_t low, int32x2_t high); // VMOV d0,d0
-int64x2_t vcombine_s64(int64x1_t low, int64x1_t high); // VMOV d0,d0
-float16x8_t vcombine_f16(float16x4_t low, float16x4_t high); // VMOV d0,d0
-float32x4_t vcombine_f32(float32x2_t low, float32x2_t high); // VMOV d0,d0
-uint8x16_t vcombine_u8(uint8x8_t low, uint8x8_t high); // VMOV d0,d0
-uint16x8_t vcombine_u16(uint16x4_t low, uint16x4_t high); // VMOV d0,d0
-uint32x4_t vcombine_u32(uint32x2_t low, uint32x2_t high); // VMOV d0,d0
-uint64x2_t vcombine_u64(uint64x1_t low, uint64x1_t high); // VMOV d0,d0
-poly8x16_t vcombine_p8(poly8x8_t low, poly8x8_t high); // VMOV d0,d0
-poly16x8_t vcombine_p16(poly16x4_t low, poly16x4_t high); // VMOV d0,d0
+_NEON2SSESTORAGE int8x16_t vcombine_s8(int8x8_t low, int8x8_t high); // VMOV d0,d0
+_NEON2SSESTORAGE int16x8_t vcombine_s16(int16x4_t low, int16x4_t high); // VMOV d0,d0
+_NEON2SSESTORAGE int32x4_t vcombine_s32(int32x2_t low, int32x2_t high); // VMOV d0,d0
+_NEON2SSESTORAGE int64x2_t vcombine_s64(int64x1_t low, int64x1_t high); // VMOV d0,d0
+_NEON2SSESTORAGE float16x8_t vcombine_f16(float16x4_t low, float16x4_t high); // VMOV d0,d0
+_NEON2SSESTORAGE float32x4_t vcombine_f32(float32x2_t low, float32x2_t high); // VMOV d0,d0
+_NEON2SSESTORAGE uint8x16_t vcombine_u8(uint8x8_t low, uint8x8_t high); // VMOV d0,d0
+_NEON2SSESTORAGE uint16x8_t vcombine_u16(uint16x4_t low, uint16x4_t high); // VMOV d0,d0
+_NEON2SSESTORAGE uint32x4_t vcombine_u32(uint32x2_t low, uint32x2_t high); // VMOV d0,d0
+_NEON2SSESTORAGE uint64x2_t vcombine_u64(uint64x1_t low, uint64x1_t high); // VMOV d0,d0
+_NEON2SSESTORAGE poly8x16_t vcombine_p8(poly8x8_t low, poly8x8_t high); // VMOV d0,d0
+_NEON2SSESTORAGE poly16x8_t vcombine_p16(poly16x4_t low, poly16x4_t high); // VMOV d0,d0
 //Splitting vectors. These intrinsics split a 128 bit vector into 2 component 64 bit vectors
-int8x8_t vget_high_s8(int8x16_t a); // VMOV d0,d0
-int16x4_t vget_high_s16(int16x8_t a); // VMOV d0,d0
-int32x2_t vget_high_s32(int32x4_t a); // VMOV d0,d0
-int64x1_t vget_high_s64(int64x2_t a); // VMOV d0,d0
-float16x4_t vget_high_f16(float16x8_t a); // VMOV d0,d0
-float32x2_t vget_high_f32(float32x4_t a); // VMOV d0,d0
-uint8x8_t vget_high_u8(uint8x16_t a); // VMOV d0,d0
-uint16x4_t vget_high_u16(uint16x8_t a); // VMOV d0,d0
-uint32x2_t vget_high_u32(uint32x4_t a); // VMOV d0,d0
-uint64x1_t vget_high_u64(uint64x2_t a); // VMOV d0,d0
-poly8x8_t vget_high_p8(poly8x16_t a); // VMOV d0,d0
-poly16x4_t vget_high_p16(poly16x8_t a); // VMOV d0,d0
-int8x8_t vget_low_s8(int8x16_t a); // VMOV d0,d0
-int16x4_t vget_low_s16(int16x8_t a); // VMOV d0,d0
-int32x2_t vget_low_s32(int32x4_t a); // VMOV d0,d0
-int64x1_t vget_low_s64(int64x2_t a); // VMOV d0,d0
-float16x4_t vget_low_f16(float16x8_t a); // VMOV d0,d0
-float32x2_t vget_low_f32(float32x4_t a); // VMOV d0,d0
-uint8x8_t vget_low_u8(uint8x16_t a); // VMOV d0,d0
-uint16x4_t vget_low_u16(uint16x8_t a); // VMOV d0,d0
-uint32x2_t vget_low_u32(uint32x4_t a); // VMOV d0,d0
-uint64x1_t vget_low_u64(uint64x2_t a); // VMOV d0,d0
-poly8x8_t vget_low_p8(poly8x16_t a); // VMOV d0,d0
-poly16x4_t vget_low_p16(poly16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int8x8_t vget_high_s8(int8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int16x4_t vget_high_s16(int16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int32x2_t vget_high_s32(int32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int64x1_t vget_high_s64(int64x2_t a); // VMOV d0,d0
+_NEON2SSESTORAGE float16x4_t vget_high_f16(float16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE float32x2_t vget_high_f32(float32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint8x8_t vget_high_u8(uint8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint16x4_t vget_high_u16(uint16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint32x2_t vget_high_u32(uint32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint64x1_t vget_high_u64(uint64x2_t a); // VMOV d0,d0
+_NEON2SSESTORAGE poly8x8_t vget_high_p8(poly8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE poly16x4_t vget_high_p16(poly16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int8x8_t vget_low_s8(int8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int16x4_t vget_low_s16(int16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int32x2_t vget_low_s32(int32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int64x1_t vget_low_s64(int64x2_t a); // VMOV d0,d0
+_NEON2SSESTORAGE float16x4_t vget_low_f16(float16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE float32x2_t vget_low_f32(float32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint8x8_t vget_low_u8(uint8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint16x4_t vget_low_u16(uint16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint32x2_t vget_low_u32(uint32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint64x1_t vget_low_u64(uint64x2_t a); // VMOV d0,d0
+_NEON2SSESTORAGE poly8x8_t vget_low_p8(poly8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE poly16x4_t vget_low_p16(poly16x8_t a); // VMOV d0,d0
 //Converting vectors. These intrinsics are used to convert vectors.
 //Convert from float
-int32x2_t vcvt_s32_f32(float32x2_t a); // VCVT.S32.F32 d0, d0
-uint32x2_t vcvt_u32_f32(float32x2_t a); // VCVT.U32.F32 d0, d0
-int32x4_t vcvtq_s32_f32(float32x4_t a); // VCVT.S32.F32 q0, q0
-uint32x4_t vcvtq_u32_f32(float32x4_t a); // VCVT.U32.F32 q0, q0
-int32x2_t vcvt_n_s32_f32(float32x2_t a, __constrange(1,32) int b); // VCVT.S32.F32 d0, d0, #32
-uint32x2_t vcvt_n_u32_f32(float32x2_t a, __constrange(1,32) int b); // VCVT.U32.F32 d0, d0, #32
-int32x4_t vcvtq_n_s32_f32(float32x4_t a, __constrange(1,32) int b); // VCVT.S32.F32 q0, q0, #32
-uint32x4_t vcvtq_n_u32_f32(float32x4_t a, __constrange(1,32) int b); // VCVT.U32.F32 q0, q0, #32
-int32x4_t vcvtnq_s32_f32(float32x4_t a); // VCVTN.S32.F32 q0, q0
+_NEON2SSESTORAGE int32x2_t vcvt_s32_f32(float32x2_t a); // VCVT.S32.F32 d0, d0
+_NEON2SSESTORAGE uint32x2_t vcvt_u32_f32(float32x2_t a); // VCVT.U32.F32 d0, d0
+_NEON2SSESTORAGE int32x4_t vcvtq_s32_f32(float32x4_t a); // VCVT.S32.F32 q0, q0
+_NEON2SSESTORAGE uint32x4_t vcvtq_u32_f32(float32x4_t a); // VCVT.U32.F32 q0, q0
+_NEON2SSESTORAGE int32x2_t vcvt_n_s32_f32(float32x2_t a, __constrange(1,32) int b); // VCVT.S32.F32 d0, d0, #32
+_NEON2SSESTORAGE uint32x2_t vcvt_n_u32_f32(float32x2_t a, __constrange(1,32) int b); // VCVT.U32.F32 d0, d0, #32
+_NEON2SSESTORAGE int32x4_t vcvtq_n_s32_f32(float32x4_t a, __constrange(1,32) int b); // VCVT.S32.F32 q0, q0, #32
+_NEON2SSESTORAGE uint32x4_t vcvtq_n_u32_f32(float32x4_t a, __constrange(1,32) int b); // VCVT.U32.F32 q0, q0, #32
+_NEON2SSESTORAGE int32x4_t vcvtnq_s32_f32(float32x4_t a); // VCVTN.S32.F32 q0, q0
 //Convert to float
-float32x2_t vcvt_f32_s32(int32x2_t a); // VCVT.F32.S32 d0, d0
-float32x2_t vcvt_f32_u32(uint32x2_t a); // VCVT.F32.U32 d0, d0
-float32x4_t vcvtq_f32_s32(int32x4_t a); // VCVT.F32.S32 q0, q0
-float32x4_t vcvtq_f32_u32(uint32x4_t a); // VCVT.F32.U32 q0, q0
-float32x2_t vcvt_n_f32_s32(int32x2_t a, __constrange(1,32) int b); // VCVT.F32.S32 d0, d0, #32
-float32x2_t vcvt_n_f32_u32(uint32x2_t a, __constrange(1,32) int b); // VCVT.F32.U32 d0, d0, #32
-float32x4_t vcvtq_n_f32_s32(int32x4_t a, __constrange(1,32) int b); // VCVT.F32.S32 q0, q0, #32
-float32x4_t vcvtq_n_f32_u32(uint32x4_t a, __constrange(1,32) int b); // VCVT.F32.U32 q0, q0, #32
+_NEON2SSESTORAGE float32x2_t vcvt_f32_s32(int32x2_t a); // VCVT.F32.S32 d0, d0
+_NEON2SSESTORAGE float32x2_t vcvt_f32_u32(uint32x2_t a); // VCVT.F32.U32 d0, d0
+_NEON2SSESTORAGE float32x4_t vcvtq_f32_s32(int32x4_t a); // VCVT.F32.S32 q0, q0
+_NEON2SSESTORAGE float32x4_t vcvtq_f32_u32(uint32x4_t a); // VCVT.F32.U32 q0, q0
+_NEON2SSESTORAGE float32x2_t vcvt_n_f32_s32(int32x2_t a, __constrange(1,32) int b); // VCVT.F32.S32 d0, d0, #32
+_NEON2SSESTORAGE float32x2_t vcvt_n_f32_u32(uint32x2_t a, __constrange(1,32) int b); // VCVT.F32.U32 d0, d0, #32
+_NEON2SSESTORAGE float32x4_t vcvtq_n_f32_s32(int32x4_t a, __constrange(1,32) int b); // VCVT.F32.S32 q0, q0, #32
+_NEON2SSESTORAGE float32x4_t vcvtq_n_f32_u32(uint32x4_t a, __constrange(1,32) int b); // VCVT.F32.U32 q0, q0, #32
 //Convert between floats
-float16x4_t vcvt_f16_f32(float32x4_t a); // VCVT.F16.F32 d0, q0
-float32x4_t vcvt_f32_f16(float16x4_t a); // VCVT.F32.F16 q0, d0
+_NEON2SSESTORAGE float16x4_t vcvt_f16_f32(float32x4_t a); // VCVT.F16.F32 d0, q0
+_NEON2SSESTORAGE float32x4_t vcvt_f32_f16(float16x4_t a); // VCVT.F32.F16 q0, d0
 //Vector narrow integer
-int8x8_t vmovn_s16(int16x8_t a); // VMOVN.I16 d0,q0
-int16x4_t vmovn_s32(int32x4_t a); // VMOVN.I32 d0,q0
-int32x2_t vmovn_s64(int64x2_t a); // VMOVN.I64 d0,q0
-uint8x8_t vmovn_u16(uint16x8_t a); // VMOVN.I16 d0,q0
-uint16x4_t vmovn_u32(uint32x4_t a); // VMOVN.I32 d0,q0
-uint32x2_t vmovn_u64(uint64x2_t a); // VMOVN.I64 d0,q0
+_NEON2SSESTORAGE int8x8_t vmovn_s16(int16x8_t a); // VMOVN.I16 d0,q0
+_NEON2SSESTORAGE int16x4_t vmovn_s32(int32x4_t a); // VMOVN.I32 d0,q0
+_NEON2SSESTORAGE int32x2_t vmovn_s64(int64x2_t a); // VMOVN.I64 d0,q0
+_NEON2SSESTORAGE uint8x8_t vmovn_u16(uint16x8_t a); // VMOVN.I16 d0,q0
+_NEON2SSESTORAGE uint16x4_t vmovn_u32(uint32x4_t a); // VMOVN.I32 d0,q0
+_NEON2SSESTORAGE uint32x2_t vmovn_u64(uint64x2_t a); // VMOVN.I64 d0,q0
 //Vector long move
-int16x8_t vmovl_s8(int8x8_t a); // VMOVL.S8 q0,d0
-int32x4_t vmovl_s16(int16x4_t a); // VMOVL.S16 q0,d0
-int64x2_t vmovl_s32(int32x2_t a); // VMOVL.S32 q0,d0
-uint16x8_t vmovl_u8(uint8x8_t a); // VMOVL.U8 q0,d0
-uint32x4_t vmovl_u16(uint16x4_t a); // VMOVL.U16 q0,d0
-uint64x2_t vmovl_u32(uint32x2_t a); // VMOVL.U32 q0,d0
+_NEON2SSESTORAGE int16x8_t vmovl_s8(int8x8_t a); // VMOVL.S8 q0,d0
+_NEON2SSESTORAGE int32x4_t vmovl_s16(int16x4_t a); // VMOVL.S16 q0,d0
+_NEON2SSESTORAGE int64x2_t vmovl_s32(int32x2_t a); // VMOVL.S32 q0,d0
+_NEON2SSESTORAGE uint16x8_t vmovl_u8(uint8x8_t a); // VMOVL.U8 q0,d0
+_NEON2SSESTORAGE uint32x4_t vmovl_u16(uint16x4_t a); // VMOVL.U16 q0,d0
+_NEON2SSESTORAGE uint64x2_t vmovl_u32(uint32x2_t a); // VMOVL.U32 q0,d0
 //Vector saturating narrow integer
-int8x8_t vqmovn_s16(int16x8_t a); // VQMOVN.S16 d0,q0
-int16x4_t vqmovn_s32(int32x4_t a); // VQMOVN.S32 d0,q0
-int32x2_t vqmovn_s64(int64x2_t a); // VQMOVN.S64 d0,q0
-uint8x8_t vqmovn_u16(uint16x8_t a); // VQMOVN.U16 d0,q0
-uint16x4_t vqmovn_u32(uint32x4_t a); // VQMOVN.U32 d0,q0
-uint32x2_t vqmovn_u64(uint64x2_t a); // VQMOVN.U64 d0,q0
+_NEON2SSESTORAGE int8x8_t vqmovn_s16(int16x8_t a); // VQMOVN.S16 d0,q0
+_NEON2SSESTORAGE int16x4_t vqmovn_s32(int32x4_t a); // VQMOVN.S32 d0,q0
+_NEON2SSESTORAGE int32x2_t vqmovn_s64(int64x2_t a); // VQMOVN.S64 d0,q0
+_NEON2SSESTORAGE uint8x8_t vqmovn_u16(uint16x8_t a); // VQMOVN.U16 d0,q0
+_NEON2SSESTORAGE uint16x4_t vqmovn_u32(uint32x4_t a); // VQMOVN.U32 d0,q0
+_NEON2SSESTORAGE uint32x2_t vqmovn_u64(uint64x2_t a); // VQMOVN.U64 d0,q0
 //Vector saturating narrow integer signed->unsigned
-uint8x8_t vqmovun_s16(int16x8_t a); // VQMOVUN.S16 d0,q0
-uint16x4_t vqmovun_s32(int32x4_t a); // VQMOVUN.S32 d0,q0
-uint32x2_t vqmovun_s64(int64x2_t a); // VQMOVUN.S64 d0,q0
+_NEON2SSESTORAGE uint8x8_t vqmovun_s16(int16x8_t a); // VQMOVUN.S16 d0,q0
+_NEON2SSESTORAGE uint16x4_t vqmovun_s32(int32x4_t a); // VQMOVUN.S32 d0,q0
+_NEON2SSESTORAGE uint32x2_t vqmovun_s64(int64x2_t a); // VQMOVUN.S64 d0,q0
 //Table look up
-uint8x8_t vtbl1_u8(uint8x8_t a, uint8x8_t b); // VTBL.8 d0, {d0}, d0
-int8x8_t vtbl1_s8(int8x8_t a, int8x8_t b); // VTBL.8 d0, {d0}, d0
-poly8x8_t vtbl1_p8(poly8x8_t a, uint8x8_t b); // VTBL.8 d0, {d0}, d0
+_NEON2SSESTORAGE uint8x8_t vtbl1_u8(uint8x8_t a, uint8x8_t b); // VTBL.8 d0, {d0}, d0
+_NEON2SSESTORAGE int8x8_t vtbl1_s8(int8x8_t a, int8x8_t b); // VTBL.8 d0, {d0}, d0
+_NEON2SSESTORAGE poly8x8_t vtbl1_p8(poly8x8_t a, uint8x8_t b); // VTBL.8 d0, {d0}, d0
 //Extended table look up intrinsics
-uint8x8_t vtbx1_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VTBX.8 d0, {d0}, d0
-int8x8_t vtbx1_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VTBX.8 d0, {d0}, d0
-poly8x8_t vtbx1_p8(poly8x8_t a, poly8x8_t b, uint8x8_t c); // VTBX.8 d0, {d0}, d0
-uint8x8_t vtbx2_u8(uint8x8_t a, uint8x8x2_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1}, d0
-int8x8_t vtbx2_s8(int8x8_t a, int8x8x2_t b, int8x8_t c); // VTBX.8 d0, {d0, d1}, d0
-poly8x8_t vtbx2_p8(poly8x8_t a, poly8x8x2_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1}, d0
-uint8x8_t vtbx3_u8(uint8x8_t a, uint8x8x3_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
-int8x8_t vtbx3_s8(int8x8_t a, int8x8x3_t b, int8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
-poly8x8_t vtbx3_p8(poly8x8_t a, poly8x8x3_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
-uint8x8_t vtbx4_u8(uint8x8_t a, uint8x8x4_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
-int8x8_t vtbx4_s8(int8x8_t a, int8x8x4_t b, int8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
-poly8x8_t vtbx4_p8(poly8x8_t a, poly8x8x4_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
+_NEON2SSESTORAGE uint8x8_t vtbx1_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VTBX.8 d0, {d0}, d0
+_NEON2SSESTORAGE int8x8_t vtbx1_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VTBX.8 d0, {d0}, d0
+_NEON2SSESTORAGE poly8x8_t vtbx1_p8(poly8x8_t a, poly8x8_t b, uint8x8_t c); // VTBX.8 d0, {d0}, d0
+_NEON2SSESTORAGE uint8x8_t vtbx2_u8(uint8x8_t a, uint8x8x2_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1}, d0
+_NEON2SSESTORAGE int8x8_t vtbx2_s8(int8x8_t a, int8x8x2_t b, int8x8_t c); // VTBX.8 d0, {d0, d1}, d0
+_NEON2SSESTORAGE poly8x8_t vtbx2_p8(poly8x8_t a, poly8x8x2_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1}, d0
+_NEON2SSESTORAGE uint8x8_t vtbx3_u8(uint8x8_t a, uint8x8x3_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
+_NEON2SSESTORAGE int8x8_t vtbx3_s8(int8x8_t a, int8x8x3_t b, int8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
+_NEON2SSESTORAGE poly8x8_t vtbx3_p8(poly8x8_t a, poly8x8x3_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
+_NEON2SSESTORAGE uint8x8_t vtbx4_u8(uint8x8_t a, uint8x8x4_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
+_NEON2SSESTORAGE int8x8_t vtbx4_s8(int8x8_t a, int8x8x4_t b, int8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
+_NEON2SSESTORAGE poly8x8_t vtbx4_p8(poly8x8_t a, poly8x8x4_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
 //Operations with a scalar value
 //Vector multiply accumulate with scalar
-int16x4_t vmla_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLA.I16 d0, d0,d0[0]
-int32x2_t vmla_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLA.I32 d0, d0,d0[0]
-uint16x4_t vmla_lane_u16(uint16x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLA.I16 d0, d0,d0[0]
-uint32x2_t vmla_lane_u32(uint32x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLA.I32 d0, d0,d0[0]
-float32x2_t vmla_lane_f32(float32x2_t a, float32x2_t b, float32x2_t v, __constrange(0,1) int l); // VMLA.F32 d0,d0, d0[0]
-int16x8_t vmlaq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v, __constrange(0,3) int l); // VMLA.I16 q0, q0,d0[0]
-int32x4_t vmlaq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v, __constrange(0,1) int l); // VMLA.I32 q0, q0,d0[0]
-uint16x8_t vmlaq_lane_u16(uint16x8_t a, uint16x8_t b, uint16x4_t v, __constrange(0,3) int l); // VMLA.I16 q0,q0, d0[0]
-uint32x4_t vmlaq_lane_u32(uint32x4_t a, uint32x4_t b, uint32x2_t v, __constrange(0,1) int l); // VMLA.I32 q0,q0, d0[0]
-float32x4_t vmlaq_lane_f32(float32x4_t a, float32x4_t b, float32x2_t v, __constrange(0,1) int l); // VMLA.F32 q0,q0, d0[0]
+_NEON2SSESTORAGE int16x4_t vmla_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLA.I16 d0, d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vmla_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLA.I32 d0, d0,d0[0]
+_NEON2SSESTORAGE uint16x4_t vmla_lane_u16(uint16x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLA.I16 d0, d0,d0[0]
+_NEON2SSESTORAGE uint32x2_t vmla_lane_u32(uint32x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLA.I32 d0, d0,d0[0]
+_NEON2SSESTORAGE float32x2_t vmla_lane_f32(float32x2_t a, float32x2_t b, float32x2_t v, __constrange(0,1) int l); // VMLA.F32 d0,d0, d0[0]
+_NEON2SSESTORAGE int16x8_t vmlaq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v, __constrange(0,3) int l); // VMLA.I16 q0, q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vmlaq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v, __constrange(0,1) int l); // VMLA.I32 q0, q0,d0[0]
+_NEON2SSESTORAGE uint16x8_t vmlaq_lane_u16(uint16x8_t a, uint16x8_t b, uint16x4_t v, __constrange(0,3) int l); // VMLA.I16 q0,q0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlaq_lane_u32(uint32x4_t a, uint32x4_t b, uint32x2_t v, __constrange(0,1) int l); // VMLA.I32 q0,q0, d0[0]
+_NEON2SSESTORAGE float32x4_t vmlaq_lane_f32(float32x4_t a, float32x4_t b, float32x2_t v, __constrange(0,1) int l); // VMLA.F32 q0,q0, d0[0]
 //Vector widening multiply accumulate with scalar
-int32x4_t vmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); //VMLAL.S16 q0, d0,d0[0]
-int64x2_t vmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); //VMLAL.S32 q0, d0,d0[0]
-uint32x4_t vmlal_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLAL.U16 q0,d0, d0[0]
-uint64x2_t vmlal_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLAL.U32 q0,d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); //VMLAL.S16 q0, d0,d0[0]
+_NEON2SSESTORAGE int64x2_t vmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); //VMLAL.S32 q0, d0,d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlal_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLAL.U16 q0,d0, d0[0]
+_NEON2SSESTORAGE uint64x2_t vmlal_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLAL.U32 q0,d0, d0[0]
 //Vector widening saturating doubling multiply accumulate with scalar
-int32x4_t vqdmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VQDMLAL.S16 q0,d0, d0[0]
-int64x2_t vqdmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VQDMLAL.S32 q0,d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VQDMLAL.S16 q0,d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VQDMLAL.S32 q0,d0, d0[0]
 //Vector multiply subtract with scalar
-int16x4_t vmls_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLS.I16 d0, d0,d0[0]
-int32x2_t vmls_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLS.I32 d0, d0,d0[0]
-uint16x4_t vmls_lane_u16(uint16x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLS.I16 d0, d0,d0[0]
-uint32x2_t vmls_lane_u32(uint32x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLS.I32 d0, d0,d0[0]
-float32x2_t vmls_lane_f32(float32x2_t a, float32x2_t b, float32x2_t v, __constrange(0,1) int l); // VMLS.F32 d0,d0, d0[0]
-int16x8_t vmlsq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v, __constrange(0,3) int l); // VMLS.I16 q0, q0,d0[0]
-int32x4_t vmlsq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v, __constrange(0,1) int l); // VMLS.I32 q0, q0,d0[0]
-uint16x8_t vmlsq_lane_u16(uint16x8_t a, uint16x8_t b, uint16x4_t v, __constrange(0,3) int l); // VMLS.I16 q0,q0, d0[0]
-uint32x4_t vmlsq_lane_u32(uint32x4_t a, uint32x4_t b, uint32x2_t v, __constrange(0,1) int l); // VMLS.I32 q0,q0, d0[0]
-float32x4_t vmlsq_lane_f32(float32x4_t a, float32x4_t b, float32x2_t v, __constrange(0,1) int l); // VMLS.F32 q0,q0, d0[0]
+_NEON2SSESTORAGE int16x4_t vmls_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLS.I16 d0, d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vmls_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLS.I32 d0, d0,d0[0]
+_NEON2SSESTORAGE uint16x4_t vmls_lane_u16(uint16x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLS.I16 d0, d0,d0[0]
+_NEON2SSESTORAGE uint32x2_t vmls_lane_u32(uint32x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLS.I32 d0, d0,d0[0]
+_NEON2SSESTORAGE float32x2_t vmls_lane_f32(float32x2_t a, float32x2_t b, float32x2_t v, __constrange(0,1) int l); // VMLS.F32 d0,d0, d0[0]
+_NEON2SSESTORAGE int16x8_t vmlsq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v, __constrange(0,3) int l); // VMLS.I16 q0, q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vmlsq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v, __constrange(0,1) int l); // VMLS.I32 q0, q0,d0[0]
+_NEON2SSESTORAGE uint16x8_t vmlsq_lane_u16(uint16x8_t a, uint16x8_t b, uint16x4_t v, __constrange(0,3) int l); // VMLS.I16 q0,q0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlsq_lane_u32(uint32x4_t a, uint32x4_t b, uint32x2_t v, __constrange(0,1) int l); // VMLS.I32 q0,q0, d0[0]
+_NEON2SSESTORAGE float32x4_t vmlsq_lane_f32(float32x4_t a, float32x4_t b, float32x2_t v, __constrange(0,1) int l); // VMLS.F32 q0,q0, d0[0]
 //Vector widening multiply subtract with scalar
-int32x4_t vmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLSL.S16 q0, d0,d0[0]
-int64x2_t vmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLSL.S32 q0, d0,d0[0]
-uint32x4_t vmlsl_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLSL.U16 q0,d0, d0[0]
-uint64x2_t vmlsl_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLSL.U32 q0,d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLSL.S16 q0, d0,d0[0]
+_NEON2SSESTORAGE int64x2_t vmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLSL.S32 q0, d0,d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlsl_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLSL.U16 q0,d0, d0[0]
+_NEON2SSESTORAGE uint64x2_t vmlsl_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLSL.U32 q0,d0, d0[0]
 //Vector widening saturating doubling multiply subtract with scalar
-int32x4_t vqdmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VQDMLSL.S16 q0,d0, d0[0]
-int64x2_t vqdmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VQDMLSL.S32 q0,d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VQDMLSL.S16 q0,d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VQDMLSL.S32 q0,d0, d0[0]
 //Vector multiply by scalar
-int16x4_t vmul_n_s16(int16x4_t a, int16_t b); // VMUL.I16 d0,d0,d0[0]
-int32x2_t vmul_n_s32(int32x2_t a, int32_t b); // VMUL.I32 d0,d0,d0[0]
-float32x2_t vmul_n_f32(float32x2_t a, float32_t b); // VMUL.F32 d0,d0,d0[0]
-uint16x4_t vmul_n_u16(uint16x4_t a, uint16_t b); // VMUL.I16 d0,d0,d0[0]
-uint32x2_t vmul_n_u32(uint32x2_t a, uint32_t b); // VMUL.I32 d0,d0,d0[0]
-int16x8_t vmulq_n_s16(int16x8_t a, int16_t b); // VMUL.I16 q0,q0,d0[0]
-int32x4_t vmulq_n_s32(int32x4_t a, int32_t b); // VMUL.I32 q0,q0,d0[0]
-float32x4_t vmulq_n_f32(float32x4_t a, float32_t b); // VMUL.F32 q0,q0,d0[0]
-uint16x8_t vmulq_n_u16(uint16x8_t a, uint16_t b); // VMUL.I16 q0,q0,d0[0]
-uint32x4_t vmulq_n_u32(uint32x4_t a, uint32_t b); // VMUL.I32 q0,q0,d0[0]
+_NEON2SSESTORAGE int16x4_t vmul_n_s16(int16x4_t a, int16_t b); // VMUL.I16 d0,d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vmul_n_s32(int32x2_t a, int32_t b); // VMUL.I32 d0,d0,d0[0]
+_NEON2SSESTORAGE float32x2_t vmul_n_f32(float32x2_t a, float32_t b); // VMUL.F32 d0,d0,d0[0]
+_NEON2SSESTORAGE uint16x4_t vmul_n_u16(uint16x4_t a, uint16_t b); // VMUL.I16 d0,d0,d0[0]
+_NEON2SSESTORAGE uint32x2_t vmul_n_u32(uint32x2_t a, uint32_t b); // VMUL.I32 d0,d0,d0[0]
+_NEON2SSESTORAGE int16x8_t vmulq_n_s16(int16x8_t a, int16_t b); // VMUL.I16 q0,q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vmulq_n_s32(int32x4_t a, int32_t b); // VMUL.I32 q0,q0,d0[0]
+_NEON2SSESTORAGE float32x4_t vmulq_n_f32(float32x4_t a, float32_t b); // VMUL.F32 q0,q0,d0[0]
+_NEON2SSESTORAGE uint16x8_t vmulq_n_u16(uint16x8_t a, uint16_t b); // VMUL.I16 q0,q0,d0[0]
+_NEON2SSESTORAGE uint32x4_t vmulq_n_u32(uint32x4_t a, uint32_t b); // VMUL.I32 q0,q0,d0[0]
 //Vector long multiply with scalar
-int32x4_t vmull_n_s16(int16x4_t vec1, int16_t val2); // VMULL.S16 q0,d0,d0[0]
-int64x2_t vmull_n_s32(int32x2_t vec1, int32_t val2); // VMULL.S32 q0,d0,d0[0]
-uint32x4_t vmull_n_u16(uint16x4_t vec1, uint16_t val2); // VMULL.U16 q0,d0,d0[0]
-uint64x2_t vmull_n_u32(uint32x2_t vec1, uint32_t val2); // VMULL.U32 q0,d0,d0[0]
+_NEON2SSESTORAGE int32x4_t vmull_n_s16(int16x4_t vec1, int16_t val2); // VMULL.S16 q0,d0,d0[0]
+_NEON2SSESTORAGE int64x2_t vmull_n_s32(int32x2_t vec1, int32_t val2); // VMULL.S32 q0,d0,d0[0]
+_NEON2SSESTORAGE uint32x4_t vmull_n_u16(uint16x4_t vec1, uint16_t val2); // VMULL.U16 q0,d0,d0[0]
+_NEON2SSESTORAGE uint64x2_t vmull_n_u32(uint32x2_t vec1, uint32_t val2); // VMULL.U32 q0,d0,d0[0]
 //Vector long multiply by scalar
-int32x4_t vmull_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VMULL.S16 q0,d0,d0[0]
-int64x2_t vmull_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VMULL.S32 q0,d0,d0[0]
-uint32x4_t vmull_lane_u16(uint16x4_t vec1, uint16x4_t val2, __constrange(0, 3) int val3); // VMULL.U16 q0,d0,d0[0]
-uint64x2_t vmull_lane_u32(uint32x2_t vec1, uint32x2_t val2, __constrange(0, 1) int val3); // VMULL.U32 q0,d0,d0[0]
+_NEON2SSESTORAGE int32x4_t vmull_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VMULL.S16 q0,d0,d0[0]
+_NEON2SSESTORAGE int64x2_t vmull_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VMULL.S32 q0,d0,d0[0]
+_NEON2SSESTORAGE uint32x4_t vmull_lane_u16(uint16x4_t vec1, uint16x4_t val2, __constrange(0, 3) int val3); // VMULL.U16 q0,d0,d0[0]
+_NEON2SSESTORAGE uint64x2_t vmull_lane_u32(uint32x2_t vec1, uint32x2_t val2, __constrange(0, 1) int val3); // VMULL.U32 q0,d0,d0[0]
 //Vector saturating doubling long multiply with scalar
-int32x4_t vqdmull_n_s16(int16x4_t vec1, int16_t val2); // VQDMULL.S16 q0,d0,d0[0]
-int64x2_t vqdmull_n_s32(int32x2_t vec1, int32_t val2); // VQDMULL.S32 q0,d0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmull_n_s16(int16x4_t vec1, int16_t val2); // VQDMULL.S16 q0,d0,d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmull_n_s32(int32x2_t vec1, int32_t val2); // VQDMULL.S32 q0,d0,d0[0]
 //Vector saturating doubling long multiply by scalar
-int32x4_t vqdmull_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQDMULL.S16 q0,d0,d0[0]
-int64x2_t vqdmull_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQDMULL.S32 q0,d0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmull_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQDMULL.S16 q0,d0,d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmull_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQDMULL.S32 q0,d0,d0[0]
 //Vector saturating doubling multiply high with scalar
-int16x4_t vqdmulh_n_s16(int16x4_t vec1, int16_t val2); // VQDMULH.S16 d0,d0,d0[0]
-int32x2_t vqdmulh_n_s32(int32x2_t vec1, int32_t val2); // VQDMULH.S32 d0,d0,d0[0]
-int16x8_t vqdmulhq_n_s16(int16x8_t vec1, int16_t val2); // VQDMULH.S16 q0,q0,d0[0]
-int32x4_t vqdmulhq_n_s32(int32x4_t vec1, int32_t val2); // VQDMULH.S32 q0,q0,d0[0]
+_NEON2SSESTORAGE int16x4_t vqdmulh_n_s16(int16x4_t vec1, int16_t val2); // VQDMULH.S16 d0,d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vqdmulh_n_s32(int32x2_t vec1, int32_t val2); // VQDMULH.S32 d0,d0,d0[0]
+_NEON2SSESTORAGE int16x8_t vqdmulhq_n_s16(int16x8_t vec1, int16_t val2); // VQDMULH.S16 q0,q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmulhq_n_s32(int32x4_t vec1, int32_t val2); // VQDMULH.S32 q0,q0,d0[0]
 //Vector saturating doubling multiply high by scalar
-int16x4_t vqdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQDMULH.S16 d0,d0,d0[0]
-int32x2_t vqdmulh_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQDMULH.S32 d0,d0,d0[0]
-int16x8_t vqdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQDMULH.S16 q0,q0,d0[0]
-int32x4_t vqdmulhq_lane_s32(int32x4_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQDMULH.S32 q0,q0,d0[0]
+_NEON2SSESTORAGE int16x4_t vqdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQDMULH.S16 d0,d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vqdmulh_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQDMULH.S32 d0,d0,d0[0]
+_NEON2SSESTORAGE int16x8_t vqdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQDMULH.S16 q0,q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmulhq_lane_s32(int32x4_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQDMULH.S32 q0,q0,d0[0]
 //Vector saturating rounding doubling multiply high with scalar
-int16x4_t vqrdmulh_n_s16(int16x4_t vec1, int16_t val2); // VQRDMULH.S16 d0,d0,d0[0]
-int32x2_t vqrdmulh_n_s32(int32x2_t vec1, int32_t val2); // VQRDMULH.S32 d0,d0,d0[0]
-int16x8_t vqrdmulhq_n_s16(int16x8_t vec1, int16_t val2); // VQRDMULH.S16 q0,q0,d0[0]
-int32x4_t vqrdmulhq_n_s32(int32x4_t vec1, int32_t val2); // VQRDMULH.S32 q0,q0,d0[0]
+_NEON2SSESTORAGE int16x4_t vqrdmulh_n_s16(int16x4_t vec1, int16_t val2); // VQRDMULH.S16 d0,d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vqrdmulh_n_s32(int32x2_t vec1, int32_t val2); // VQRDMULH.S32 d0,d0,d0[0]
+_NEON2SSESTORAGE int16x8_t vqrdmulhq_n_s16(int16x8_t vec1, int16_t val2); // VQRDMULH.S16 q0,q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqrdmulhq_n_s32(int32x4_t vec1, int32_t val2); // VQRDMULH.S32 q0,q0,d0[0]
 //Vector rounding saturating doubling multiply high by scalar
-int16x4_t vqrdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQRDMULH.S16 d0,d0,d0[0]
-int32x2_t vqrdmulh_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQRDMULH.S32 d0,d0,d0[0]
-int16x8_t vqrdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQRDMULH.S16 q0,q0,d0[0]
-int32x4_t vqrdmulhq_lane_s32(int32x4_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQRDMULH.S32 q0,q0,d0[0]
+_NEON2SSESTORAGE int16x4_t vqrdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQRDMULH.S16 d0,d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vqrdmulh_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQRDMULH.S32 d0,d0,d0[0]
+_NEON2SSESTORAGE int16x8_t vqrdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQRDMULH.S16 q0,q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqrdmulhq_lane_s32(int32x4_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQRDMULH.S32 q0,q0,d0[0]
 //Vector multiply accumulate with scalar
-int16x4_t vmla_n_s16(int16x4_t a, int16x4_t b, int16_t c); // VMLA.I16 d0, d0, d0[0]
-int32x2_t vmla_n_s32(int32x2_t a, int32x2_t b, int32_t c); // VMLA.I32 d0, d0, d0[0]
-uint16x4_t vmla_n_u16(uint16x4_t a, uint16x4_t b, uint16_t c); // VMLA.I16 d0, d0, d0[0]
-uint32x2_t vmla_n_u32(uint32x2_t a, uint32x2_t b, uint32_t c); // VMLA.I32 d0, d0, d0[0]
-float32x2_t vmla_n_f32(float32x2_t a, float32x2_t b, float32_t c); // VMLA.F32 d0, d0, d0[0]
-int16x8_t vmlaq_n_s16(int16x8_t a, int16x8_t b, int16_t c); // VMLA.I16 q0, q0, d0[0]
-int32x4_t vmlaq_n_s32(int32x4_t a, int32x4_t b, int32_t c); // VMLA.I32 q0, q0, d0[0]
-uint16x8_t vmlaq_n_u16(uint16x8_t a, uint16x8_t b, uint16_t c); // VMLA.I16 q0, q0, d0[0]
-uint32x4_t vmlaq_n_u32(uint32x4_t a, uint32x4_t b, uint32_t c); // VMLA.I32 q0, q0, d0[0]
-float32x4_t vmlaq_n_f32(float32x4_t a, float32x4_t b, float32_t c); // VMLA.F32 q0, q0, d0[0]
+_NEON2SSESTORAGE int16x4_t vmla_n_s16(int16x4_t a, int16x4_t b, int16_t c); // VMLA.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE int32x2_t vmla_n_s32(int32x2_t a, int32x2_t b, int32_t c); // VMLA.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE uint16x4_t vmla_n_u16(uint16x4_t a, uint16x4_t b, uint16_t c); // VMLA.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE uint32x2_t vmla_n_u32(uint32x2_t a, uint32x2_t b, uint32_t c); // VMLA.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE float32x2_t vmla_n_f32(float32x2_t a, float32x2_t b, float32_t c); // VMLA.F32 d0, d0, d0[0]
+_NEON2SSESTORAGE int16x8_t vmlaq_n_s16(int16x8_t a, int16x8_t b, int16_t c); // VMLA.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlaq_n_s32(int32x4_t a, int32x4_t b, int32_t c); // VMLA.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE uint16x8_t vmlaq_n_u16(uint16x8_t a, uint16x8_t b, uint16_t c); // VMLA.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlaq_n_u32(uint32x4_t a, uint32x4_t b, uint32_t c); // VMLA.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE float32x4_t vmlaq_n_f32(float32x4_t a, float32x4_t b, float32_t c); // VMLA.F32 q0, q0, d0[0]
 //Vector widening multiply accumulate with scalar
-int32x4_t vmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VMLAL.S16 q0, d0, d0[0]
-int64x2_t vmlal_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VMLAL.S32 q0, d0, d0[0]
-uint32x4_t vmlal_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c); // VMLAL.U16 q0, d0, d0[0]
-uint64x2_t vmlal_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c); // VMLAL.U32 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VMLAL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vmlal_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VMLAL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlal_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c); // VMLAL.U16 q0, d0, d0[0]
+_NEON2SSESTORAGE uint64x2_t vmlal_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c); // VMLAL.U32 q0, d0, d0[0]
 //Vector widening saturating doubling multiply accumulate with scalar
-int32x4_t vqdmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VQDMLAL.S16 q0, d0, d0[0]
-int64x2_t vqdmlal_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VQDMLAL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VQDMLAL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmlal_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VQDMLAL.S32 q0, d0, d0[0]
 //Vector multiply subtract with scalar
-int16x4_t vmls_n_s16(int16x4_t a, int16x4_t b, int16_t c); // VMLS.I16 d0, d0, d0[0]
-int32x2_t vmls_n_s32(int32x2_t a, int32x2_t b, int32_t c); // VMLS.I32 d0, d0, d0[0]
-uint16x4_t vmls_n_u16(uint16x4_t a, uint16x4_t b, uint16_t c); // VMLS.I16 d0, d0, d0[0]
-uint32x2_t vmls_n_u32(uint32x2_t a, uint32x2_t b, uint32_t c); // VMLS.I32 d0, d0, d0[0]
-float32x2_t vmls_n_f32(float32x2_t a, float32x2_t b, float32_t c); // VMLS.F32 d0, d0, d0[0]
-int16x8_t vmlsq_n_s16(int16x8_t a, int16x8_t b, int16_t c); // VMLS.I16 q0, q0, d0[0]
-int32x4_t vmlsq_n_s32(int32x4_t a, int32x4_t b, int32_t c); // VMLS.I32 q0, q0, d0[0]
-uint16x8_t vmlsq_n_u16(uint16x8_t a, uint16x8_t b, uint16_t c); // VMLS.I16 q0, q0, d0[0]
-uint32x4_t vmlsq_n_u32(uint32x4_t a, uint32x4_t b, uint32_t c); // VMLS.I32 q0, q0, d0[0]
-float32x4_t vmlsq_n_f32(float32x4_t a, float32x4_t b, float32_t c); // VMLS.F32 q0, q0, d0[0]
+_NEON2SSESTORAGE int16x4_t vmls_n_s16(int16x4_t a, int16x4_t b, int16_t c); // VMLS.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE int32x2_t vmls_n_s32(int32x2_t a, int32x2_t b, int32_t c); // VMLS.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE uint16x4_t vmls_n_u16(uint16x4_t a, uint16x4_t b, uint16_t c); // VMLS.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE uint32x2_t vmls_n_u32(uint32x2_t a, uint32x2_t b, uint32_t c); // VMLS.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE float32x2_t vmls_n_f32(float32x2_t a, float32x2_t b, float32_t c); // VMLS.F32 d0, d0, d0[0]
+_NEON2SSESTORAGE int16x8_t vmlsq_n_s16(int16x8_t a, int16x8_t b, int16_t c); // VMLS.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlsq_n_s32(int32x4_t a, int32x4_t b, int32_t c); // VMLS.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE uint16x8_t vmlsq_n_u16(uint16x8_t a, uint16x8_t b, uint16_t c); // VMLS.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlsq_n_u32(uint32x4_t a, uint32x4_t b, uint32_t c); // VMLS.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE float32x4_t vmlsq_n_f32(float32x4_t a, float32x4_t b, float32_t c); // VMLS.F32 q0, q0, d0[0]
 //Vector widening multiply subtract with scalar
-int32x4_t vmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VMLSL.S16 q0, d0, d0[0]
-int64x2_t vmlsl_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VMLSL.S32 q0, d0, d0[0]
-uint32x4_t vmlsl_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c); // VMLSL.U16 q0, d0, d0[0]
-uint64x2_t vmlsl_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c); // VMLSL.U32 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VMLSL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vmlsl_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VMLSL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlsl_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c); // VMLSL.U16 q0, d0, d0[0]
+_NEON2SSESTORAGE uint64x2_t vmlsl_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c); // VMLSL.U32 q0, d0, d0[0]
 //Vector widening saturating doubling multiply subtract with scalar
-int32x4_t vqdmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VQDMLSL.S16 q0, d0, d0[0]
-int64x2_t vqdmlsl_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VQDMLSL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VQDMLSL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmlsl_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VQDMLSL.S32 q0, d0, d0[0]
 //Vector extract
-int8x8_t vext_s8(int8x8_t a, int8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
-uint8x8_t vext_u8(uint8x8_t a, uint8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
-poly8x8_t vext_p8(poly8x8_t a, poly8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
-int16x4_t vext_s16(int16x4_t a, int16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
-uint16x4_t vext_u16(uint16x4_t a, uint16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
-poly16x4_t vext_p16(poly16x4_t a, poly16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
-int32x2_t vext_s32(int32x2_t a, int32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
-uint32x2_t vext_u32(uint32x2_t a, uint32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
-int64x1_t vext_s64(int64x1_t a, int64x1_t b, __constrange(0,0) int c); // VEXT.64 d0,d0,d0,#0
-uint64x1_t vext_u64(uint64x1_t a, uint64x1_t b, __constrange(0,0) int c); // VEXT.64 d0,d0,d0,#0
-float32x2_t vext_f32(float32x2_t a, float32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
-int8x16_t vextq_s8(int8x16_t a, int8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
-uint8x16_t vextq_u8(uint8x16_t a, uint8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
-poly8x16_t vextq_p8(poly8x16_t a, poly8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
-int16x8_t vextq_s16(int16x8_t a, int16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
-uint16x8_t vextq_u16(uint16x8_t a, uint16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
-poly16x8_t vextq_p16(poly16x8_t a, poly16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
-int32x4_t vextq_s32(int32x4_t a, int32x4_t b, __constrange(0,3) int c); // VEXT.32 q0,q0,q0,#0
-uint32x4_t vextq_u32(uint32x4_t a, uint32x4_t b, __constrange(0,3) int c); // VEXT.32 q0,q0,q0,#0
-int64x2_t vextq_s64(int64x2_t a, int64x2_t b, __constrange(0,1) int c); // VEXT.64 q0,q0,q0,#0
-uint64x2_t vextq_u64(uint64x2_t a, uint64x2_t b, __constrange(0,1) int c); // VEXT.64 q0,q0,q0,#0
-float32x4_t vextq_f32(float32x4_t a, float32x4_t b, __constrange(0,3) float c); // VEXT.32 q0,q0,q0,#0
+_NEON2SSESTORAGE int8x8_t vext_s8(int8x8_t a, int8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
+_NEON2SSESTORAGE uint8x8_t vext_u8(uint8x8_t a, uint8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
+_NEON2SSESTORAGE poly8x8_t vext_p8(poly8x8_t a, poly8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
+_NEON2SSESTORAGE int16x4_t vext_s16(int16x4_t a, int16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
+_NEON2SSESTORAGE uint16x4_t vext_u16(uint16x4_t a, uint16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
+_NEON2SSESTORAGE poly16x4_t vext_p16(poly16x4_t a, poly16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
+_NEON2SSESTORAGE int32x2_t vext_s32(int32x2_t a, int32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
+_NEON2SSESTORAGE uint32x2_t vext_u32(uint32x2_t a, uint32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
+_NEON2SSESTORAGE int64x1_t vext_s64(int64x1_t a, int64x1_t b, __constrange(0,0) int c); // VEXT.64 d0,d0,d0,#0
+_NEON2SSESTORAGE uint64x1_t vext_u64(uint64x1_t a, uint64x1_t b, __constrange(0,0) int c); // VEXT.64 d0,d0,d0,#0
+_NEON2SSESTORAGE float32x2_t vext_f32(float32x2_t a, float32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
+_NEON2SSESTORAGE int8x16_t vextq_s8(int8x16_t a, int8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
+_NEON2SSESTORAGE uint8x16_t vextq_u8(uint8x16_t a, uint8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
+_NEON2SSESTORAGE poly8x16_t vextq_p8(poly8x16_t a, poly8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
+_NEON2SSESTORAGE int16x8_t vextq_s16(int16x8_t a, int16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
+_NEON2SSESTORAGE uint16x8_t vextq_u16(uint16x8_t a, uint16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
+_NEON2SSESTORAGE poly16x8_t vextq_p16(poly16x8_t a, poly16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
+_NEON2SSESTORAGE int32x4_t vextq_s32(int32x4_t a, int32x4_t b, __constrange(0,3) int c); // VEXT.32 q0,q0,q0,#0
+_NEON2SSESTORAGE uint32x4_t vextq_u32(uint32x4_t a, uint32x4_t b, __constrange(0,3) int c); // VEXT.32 q0,q0,q0,#0
+_NEON2SSESTORAGE int64x2_t vextq_s64(int64x2_t a, int64x2_t b, __constrange(0,1) int c); // VEXT.64 q0,q0,q0,#0
+_NEON2SSESTORAGE uint64x2_t vextq_u64(uint64x2_t a, uint64x2_t b, __constrange(0,1) int c); // VEXT.64 q0,q0,q0,#0
+_NEON2SSESTORAGE float32x4_t vextq_f32(float32x4_t a, float32x4_t b, __constrange(0,3) float c); // VEXT.32 q0,q0,q0,#0
 //Reverse vector elements (swap endianness). VREVn.m reverses the order of the m-bit lanes within a set that is n bits wide.
-int8x8_t vrev64_s8(int8x8_t vec); // VREV64.8 d0,d0
-int16x4_t vrev64_s16(int16x4_t vec); // VREV64.16 d0,d0
-int32x2_t vrev64_s32(int32x2_t vec); // VREV64.32 d0,d0
-uint8x8_t vrev64_u8(uint8x8_t vec); // VREV64.8 d0,d0
-uint16x4_t vrev64_u16(uint16x4_t vec); // VREV64.16 d0,d0
-uint32x2_t vrev64_u32(uint32x2_t vec); // VREV64.32 d0,d0
-poly8x8_t vrev64_p8(poly8x8_t vec); // VREV64.8 d0,d0
-poly16x4_t vrev64_p16(poly16x4_t vec); // VREV64.16 d0,d0
-float32x2_t vrev64_f32(float32x2_t vec); // VREV64.32 d0,d0
-int8x16_t vrev64q_s8(int8x16_t vec); // VREV64.8 q0,q0
-int16x8_t vrev64q_s16(int16x8_t vec); // VREV64.16 q0,q0
-int32x4_t vrev64q_s32(int32x4_t vec); // VREV64.32 q0,q0
-uint8x16_t vrev64q_u8(uint8x16_t vec); // VREV64.8 q0,q0
-uint16x8_t vrev64q_u16(uint16x8_t vec); // VREV64.16 q0,q0
-uint32x4_t vrev64q_u32(uint32x4_t vec); // VREV64.32 q0,q0
-poly8x16_t vrev64q_p8(poly8x16_t vec); // VREV64.8 q0,q0
-poly16x8_t vrev64q_p16(poly16x8_t vec); // VREV64.16 q0,q0
-float32x4_t vrev64q_f32(float32x4_t vec); // VREV64.32 q0,q0
-int8x8_t vrev32_s8(int8x8_t vec); // VREV32.8 d0,d0
-int16x4_t vrev32_s16(int16x4_t vec); // VREV32.16 d0,d0
-uint8x8_t vrev32_u8(uint8x8_t vec); // VREV32.8 d0,d0
-uint16x4_t vrev32_u16(uint16x4_t vec); // VREV32.16 d0,d0
-poly8x8_t vrev32_p8(poly8x8_t vec); // VREV32.8 d0,d0
-poly16x4_t vrev32_p16(poly16x4_t vec); // VREV32.16 d0,d0
-int8x16_t vrev32q_s8(int8x16_t vec); // VREV32.8 q0,q0
-int16x8_t vrev32q_s16(int16x8_t vec); // VREV32.16 q0,q0
-uint8x16_t vrev32q_u8(uint8x16_t vec); // VREV32.8 q0,q0
-uint16x8_t vrev32q_u16(uint16x8_t vec); // VREV32.16 q0,q0
-poly8x16_t vrev32q_p8(poly8x16_t vec); // VREV32.8 q0,q0
-poly16x8_t vrev32q_p16(poly16x8_t vec); // VREV32.16 q0,q0
-int8x8_t vrev16_s8(int8x8_t vec); // VREV16.8 d0,d0
-uint8x8_t vrev16_u8(uint8x8_t vec); // VREV16.8 d0,d0
-poly8x8_t vrev16_p8(poly8x8_t vec); // VREV16.8 d0,d0
-int8x16_t vrev16q_s8(int8x16_t vec); // VREV16.8 q0,q0
-uint8x16_t vrev16q_u8(uint8x16_t vec); // VREV16.8 q0,q0
-poly8x16_t vrev16q_p8(poly8x16_t vec); // VREV16.8 q0,q0
+_NEON2SSESTORAGE int8x8_t vrev64_s8(int8x8_t vec); // VREV64.8 d0,d0
+_NEON2SSESTORAGE int16x4_t vrev64_s16(int16x4_t vec); // VREV64.16 d0,d0
+_NEON2SSESTORAGE int32x2_t vrev64_s32(int32x2_t vec); // VREV64.32 d0,d0
+_NEON2SSESTORAGE uint8x8_t vrev64_u8(uint8x8_t vec); // VREV64.8 d0,d0
+_NEON2SSESTORAGE uint16x4_t vrev64_u16(uint16x4_t vec); // VREV64.16 d0,d0
+_NEON2SSESTORAGE uint32x2_t vrev64_u32(uint32x2_t vec); // VREV64.32 d0,d0
+_NEON2SSESTORAGE poly8x8_t vrev64_p8(poly8x8_t vec); // VREV64.8 d0,d0
+_NEON2SSESTORAGE poly16x4_t vrev64_p16(poly16x4_t vec); // VREV64.16 d0,d0
+_NEON2SSESTORAGE float32x2_t vrev64_f32(float32x2_t vec); // VREV64.32 d0,d0
+_NEON2SSESTORAGE int8x16_t vrev64q_s8(int8x16_t vec); // VREV64.8 q0,q0
+_NEON2SSESTORAGE int16x8_t vrev64q_s16(int16x8_t vec); // VREV64.16 q0,q0
+_NEON2SSESTORAGE int32x4_t vrev64q_s32(int32x4_t vec); // VREV64.32 q0,q0
+_NEON2SSESTORAGE uint8x16_t vrev64q_u8(uint8x16_t vec); // VREV64.8 q0,q0
+_NEON2SSESTORAGE uint16x8_t vrev64q_u16(uint16x8_t vec); // VREV64.16 q0,q0
+_NEON2SSESTORAGE uint32x4_t vrev64q_u32(uint32x4_t vec); // VREV64.32 q0,q0
+_NEON2SSESTORAGE poly8x16_t vrev64q_p8(poly8x16_t vec); // VREV64.8 q0,q0
+_NEON2SSESTORAGE poly16x8_t vrev64q_p16(poly16x8_t vec); // VREV64.16 q0,q0
+_NEON2SSESTORAGE float32x4_t vrev64q_f32(float32x4_t vec); // VREV64.32 q0,q0
+_NEON2SSESTORAGE int8x8_t vrev32_s8(int8x8_t vec); // VREV32.8 d0,d0
+_NEON2SSESTORAGE int16x4_t vrev32_s16(int16x4_t vec); // VREV32.16 d0,d0
+_NEON2SSESTORAGE uint8x8_t vrev32_u8(uint8x8_t vec); // VREV32.8 d0,d0
+_NEON2SSESTORAGE uint16x4_t vrev32_u16(uint16x4_t vec); // VREV32.16 d0,d0
+_NEON2SSESTORAGE poly8x8_t vrev32_p8(poly8x8_t vec); // VREV32.8 d0,d0
+_NEON2SSESTORAGE poly16x4_t vrev32_p16(poly16x4_t vec); // VREV32.16 d0,d0
+_NEON2SSESTORAGE int8x16_t vrev32q_s8(int8x16_t vec); // VREV32.8 q0,q0
+_NEON2SSESTORAGE int16x8_t vrev32q_s16(int16x8_t vec); // VREV32.16 q0,q0
+_NEON2SSESTORAGE uint8x16_t vrev32q_u8(uint8x16_t vec); // VREV32.8 q0,q0
+_NEON2SSESTORAGE uint16x8_t vrev32q_u16(uint16x8_t vec); // VREV32.16 q0,q0
+_NEON2SSESTORAGE poly8x16_t vrev32q_p8(poly8x16_t vec); // VREV32.8 q0,q0
+_NEON2SSESTORAGE poly16x8_t vrev32q_p16(poly16x8_t vec); // VREV32.16 q0,q0
+_NEON2SSESTORAGE int8x8_t vrev16_s8(int8x8_t vec); // VREV16.8 d0,d0
+_NEON2SSESTORAGE uint8x8_t vrev16_u8(uint8x8_t vec); // VREV16.8 d0,d0
+_NEON2SSESTORAGE poly8x8_t vrev16_p8(poly8x8_t vec); // VREV16.8 d0,d0
+_NEON2SSESTORAGE int8x16_t vrev16q_s8(int8x16_t vec); // VREV16.8 q0,q0
+_NEON2SSESTORAGE uint8x16_t vrev16q_u8(uint8x16_t vec); // VREV16.8 q0,q0
+_NEON2SSESTORAGE poly8x16_t vrev16q_p8(poly8x16_t vec); // VREV16.8 q0,q0
 //Other single operand arithmetic
 //Absolute: Vd[i] = |Va[i]|
-int8x8_t vabs_s8(int8x8_t a); // VABS.S8 d0,d0
-int16x4_t vabs_s16(int16x4_t a); // VABS.S16 d0,d0
-int32x2_t vabs_s32(int32x2_t a); // VABS.S32 d0,d0
-float32x2_t vabs_f32(float32x2_t a); // VABS.F32 d0,d0
-int8x16_t vabsq_s8(int8x16_t a); // VABS.S8 q0,q0
-int16x8_t vabsq_s16(int16x8_t a); // VABS.S16 q0,q0
-int32x4_t vabsq_s32(int32x4_t a); // VABS.S32 q0,q0
-float32x4_t vabsq_f32(float32x4_t a); // VABS.F32 q0,q0
-
-int64x2_t vabsq_s64(int64x2_t a); // VABS.S64 q0,q0
-float64x2_t vabsq_f64(float64x2_t a); // VABS.F64 q0,q0
+_NEON2SSESTORAGE int8x8_t vabs_s8(int8x8_t a); // VABS.S8 d0,d0
+_NEON2SSESTORAGE int16x4_t vabs_s16(int16x4_t a); // VABS.S16 d0,d0
+_NEON2SSESTORAGE int32x2_t vabs_s32(int32x2_t a); // VABS.S32 d0,d0
+_NEON2SSESTORAGE float32x2_t vabs_f32(float32x2_t a); // VABS.F32 d0,d0
+_NEON2SSESTORAGE int8x16_t vabsq_s8(int8x16_t a); // VABS.S8 q0,q0
+_NEON2SSESTORAGE int16x8_t vabsq_s16(int16x8_t a); // VABS.S16 q0,q0
+_NEON2SSESTORAGE int32x4_t vabsq_s32(int32x4_t a); // VABS.S32 q0,q0
+_NEON2SSESTORAGE float32x4_t vabsq_f32(float32x4_t a); // VABS.F32 q0,q0
+
+_NEON2SSESTORAGE int64x2_t vabsq_s64(int64x2_t a); // VABS.S64 q0,q0
+_NEON2SSESTORAGE float64x2_t vabsq_f64(float64x2_t a); // VABS.F64 q0,q0
 
 //Saturating absolute: Vd[i] = sat(|Va[i]|)
-int8x8_t vqabs_s8(int8x8_t a); // VQABS.S8 d0,d0
-int16x4_t vqabs_s16(int16x4_t a); // VQABS.S16 d0,d0
-int32x2_t vqabs_s32(int32x2_t a); // VQABS.S32 d0,d0
-int8x16_t vqabsq_s8(int8x16_t a); // VQABS.S8 q0,q0
-int16x8_t vqabsq_s16(int16x8_t a); // VQABS.S16 q0,q0
-int32x4_t vqabsq_s32(int32x4_t a); // VQABS.S32 q0,q0
+_NEON2SSESTORAGE int8x8_t vqabs_s8(int8x8_t a); // VQABS.S8 d0,d0
+_NEON2SSESTORAGE int16x4_t vqabs_s16(int16x4_t a); // VQABS.S16 d0,d0
+_NEON2SSESTORAGE int32x2_t vqabs_s32(int32x2_t a); // VQABS.S32 d0,d0
+_NEON2SSESTORAGE int8x16_t vqabsq_s8(int8x16_t a); // VQABS.S8 q0,q0
+_NEON2SSESTORAGE int16x8_t vqabsq_s16(int16x8_t a); // VQABS.S16 q0,q0
+_NEON2SSESTORAGE int32x4_t vqabsq_s32(int32x4_t a); // VQABS.S32 q0,q0
 //Negate: Vd[i] = - Va[i]
-int8x8_t vneg_s8(int8x8_t a); // VNE//d0,d0
-int16x4_t vneg_s16(int16x4_t a); // VNE//d0,d0
-int32x2_t vneg_s32(int32x2_t a); // VNE//d0,d0
-float32x2_t vneg_f32(float32x2_t a); // VNE//d0,d0
-int8x16_t vnegq_s8(int8x16_t a); // VNE//q0,q0
-int16x8_t vnegq_s16(int16x8_t a); // VNE//q0,q0
-int32x4_t vnegq_s32(int32x4_t a); // VNE//q0,q0
-float32x4_t vnegq_f32(float32x4_t a); // VNE//q0,q0
+_NEON2SSESTORAGE int8x8_t vneg_s8(int8x8_t a); // VNE//d0,d0
+_NEON2SSESTORAGE int16x4_t vneg_s16(int16x4_t a); // VNE//d0,d0
+_NEON2SSESTORAGE int32x2_t vneg_s32(int32x2_t a); // VNE//d0,d0
+_NEON2SSESTORAGE float32x2_t vneg_f32(float32x2_t a); // VNE//d0,d0
+_NEON2SSESTORAGE int8x16_t vnegq_s8(int8x16_t a); // VNE//q0,q0
+_NEON2SSESTORAGE int16x8_t vnegq_s16(int16x8_t a); // VNE//q0,q0
+_NEON2SSESTORAGE int32x4_t vnegq_s32(int32x4_t a); // VNE//q0,q0
+_NEON2SSESTORAGE float32x4_t vnegq_f32(float32x4_t a); // VNE//q0,q0
 //Saturating Negate: sat(Vd[i] = - Va[i])
-int8x8_t vqneg_s8(int8x8_t a); // VQNE//d0,d0
-int16x4_t vqneg_s16(int16x4_t a); // VQNE//d0,d0
-int32x2_t vqneg_s32(int32x2_t a); // VQNE//d0,d0
-int8x16_t vqnegq_s8(int8x16_t a); // VQNE//q0,q0
-int16x8_t vqnegq_s16(int16x8_t a); // VQNE//q0,q0
-int32x4_t vqnegq_s32(int32x4_t a); // VQNE//q0,q0
+_NEON2SSESTORAGE int8x8_t vqneg_s8(int8x8_t a); // VQNE//d0,d0
+_NEON2SSESTORAGE int16x4_t vqneg_s16(int16x4_t a); // VQNE//d0,d0
+_NEON2SSESTORAGE int32x2_t vqneg_s32(int32x2_t a); // VQNE//d0,d0
+_NEON2SSESTORAGE int8x16_t vqnegq_s8(int8x16_t a); // VQNE//q0,q0
+_NEON2SSESTORAGE int16x8_t vqnegq_s16(int16x8_t a); // VQNE//q0,q0
+_NEON2SSESTORAGE int32x4_t vqnegq_s32(int32x4_t a); // VQNE//q0,q0
 //Count leading sign bits
-int8x8_t vcls_s8(int8x8_t a); // VCLS.S8 d0,d0
-int16x4_t vcls_s16(int16x4_t a); // VCLS.S16 d0,d0
-int32x2_t vcls_s32(int32x2_t a); // VCLS.S32 d0,d0
-int8x16_t vclsq_s8(int8x16_t a); // VCLS.S8 q0,q0
-int16x8_t vclsq_s16(int16x8_t a); // VCLS.S16 q0,q0
-int32x4_t vclsq_s32(int32x4_t a); // VCLS.S32 q0,q0
+_NEON2SSESTORAGE int8x8_t vcls_s8(int8x8_t a); // VCLS.S8 d0,d0
+_NEON2SSESTORAGE int16x4_t vcls_s16(int16x4_t a); // VCLS.S16 d0,d0
+_NEON2SSESTORAGE int32x2_t vcls_s32(int32x2_t a); // VCLS.S32 d0,d0
+_NEON2SSESTORAGE int8x16_t vclsq_s8(int8x16_t a); // VCLS.S8 q0,q0
+_NEON2SSESTORAGE int16x8_t vclsq_s16(int16x8_t a); // VCLS.S16 q0,q0
+_NEON2SSESTORAGE int32x4_t vclsq_s32(int32x4_t a); // VCLS.S32 q0,q0
 //Count leading zeros
-int8x8_t vclz_s8(int8x8_t a); // VCLZ.I8 d0,d0
-int16x4_t vclz_s16(int16x4_t a); // VCLZ.I16 d0,d0
-int32x2_t vclz_s32(int32x2_t a); // VCLZ.I32 d0,d0
-uint8x8_t vclz_u8(uint8x8_t a); // VCLZ.I8 d0,d0
-uint16x4_t vclz_u16(uint16x4_t a); // VCLZ.I16 d0,d0
-uint32x2_t vclz_u32(uint32x2_t a); // VCLZ.I32 d0,d0
-int8x16_t vclzq_s8(int8x16_t a); // VCLZ.I8 q0,q0
-int16x8_t vclzq_s16(int16x8_t a); // VCLZ.I16 q0,q0
-int32x4_t vclzq_s32(int32x4_t a); // VCLZ.I32 q0,q0
-uint8x16_t vclzq_u8(uint8x16_t a); // VCLZ.I8 q0,q0
-uint16x8_t vclzq_u16(uint16x8_t a); // VCLZ.I16 q0,q0
-uint32x4_t vclzq_u32(uint32x4_t a); // VCLZ.I32 q0,q0
+_NEON2SSESTORAGE int8x8_t vclz_s8(int8x8_t a); // VCLZ.I8 d0,d0
+_NEON2SSESTORAGE int16x4_t vclz_s16(int16x4_t a); // VCLZ.I16 d0,d0
+_NEON2SSESTORAGE int32x2_t vclz_s32(int32x2_t a); // VCLZ.I32 d0,d0
+_NEON2SSESTORAGE uint8x8_t vclz_u8(uint8x8_t a); // VCLZ.I8 d0,d0
+_NEON2SSESTORAGE uint16x4_t vclz_u16(uint16x4_t a); // VCLZ.I16 d0,d0
+_NEON2SSESTORAGE uint32x2_t vclz_u32(uint32x2_t a); // VCLZ.I32 d0,d0
+_NEON2SSESTORAGE int8x16_t vclzq_s8(int8x16_t a); // VCLZ.I8 q0,q0
+_NEON2SSESTORAGE int16x8_t vclzq_s16(int16x8_t a); // VCLZ.I16 q0,q0
+_NEON2SSESTORAGE int32x4_t vclzq_s32(int32x4_t a); // VCLZ.I32 q0,q0
+_NEON2SSESTORAGE uint8x16_t vclzq_u8(uint8x16_t a); // VCLZ.I8 q0,q0
+_NEON2SSESTORAGE uint16x8_t vclzq_u16(uint16x8_t a); // VCLZ.I16 q0,q0
+_NEON2SSESTORAGE uint32x4_t vclzq_u32(uint32x4_t a); // VCLZ.I32 q0,q0
 //Count number of set bits
-uint8x8_t vcnt_u8(uint8x8_t a); // VCNT.8 d0,d0
-int8x8_t vcnt_s8(int8x8_t a); // VCNT.8 d0,d0
-poly8x8_t vcnt_p8(poly8x8_t a); // VCNT.8 d0,d0
-uint8x16_t vcntq_u8(uint8x16_t a); // VCNT.8 q0,q0
-int8x16_t vcntq_s8(int8x16_t a); // VCNT.8 q0,q0
-poly8x16_t vcntq_p8(poly8x16_t a); // VCNT.8 q0,q0
+_NEON2SSESTORAGE uint8x8_t vcnt_u8(uint8x8_t a); // VCNT.8 d0,d0
+_NEON2SSESTORAGE int8x8_t vcnt_s8(int8x8_t a); // VCNT.8 d0,d0
+_NEON2SSESTORAGE poly8x8_t vcnt_p8(poly8x8_t a); // VCNT.8 d0,d0
+_NEON2SSESTORAGE uint8x16_t vcntq_u8(uint8x16_t a); // VCNT.8 q0,q0
+_NEON2SSESTORAGE int8x16_t vcntq_s8(int8x16_t a); // VCNT.8 q0,q0
+_NEON2SSESTORAGE poly8x16_t vcntq_p8(poly8x16_t a); // VCNT.8 q0,q0
 //Reciprocal estimate
-float32x2_t vrecpe_f32(float32x2_t a); // VRECPE.F32 d0,d0
-uint32x2_t vrecpe_u32(uint32x2_t a); // VRECPE.U32 d0,d0
-float32x4_t vrecpeq_f32(float32x4_t a); // VRECPE.F32 q0,q0
-uint32x4_t vrecpeq_u32(uint32x4_t a); // VRECPE.U32 q0,q0
+_NEON2SSESTORAGE float32x2_t vrecpe_f32(float32x2_t a); // VRECPE.F32 d0,d0
+_NEON2SSESTORAGE uint32x2_t vrecpe_u32(uint32x2_t a); // VRECPE.U32 d0,d0
+_NEON2SSESTORAGE float32x4_t vrecpeq_f32(float32x4_t a); // VRECPE.F32 q0,q0
+_NEON2SSESTORAGE uint32x4_t vrecpeq_u32(uint32x4_t a); // VRECPE.U32 q0,q0
 //Reciprocal square root estimate
-float32x2_t vrsqrte_f32(float32x2_t a); // VRSQRTE.F32 d0,d0
-uint32x2_t vrsqrte_u32(uint32x2_t a); // VRSQRTE.U32 d0,d0
-float32x4_t vrsqrteq_f32(float32x4_t a); // VRSQRTE.F32 q0,q0
-uint32x4_t vrsqrteq_u32(uint32x4_t a); // VRSQRTE.U32 q0,q0
+_NEON2SSESTORAGE float32x2_t vrsqrte_f32(float32x2_t a); // VRSQRTE.F32 d0,d0
+_NEON2SSESTORAGE uint32x2_t vrsqrte_u32(uint32x2_t a); // VRSQRTE.U32 d0,d0
+_NEON2SSESTORAGE float32x4_t vrsqrteq_f32(float32x4_t a); // VRSQRTE.F32 q0,q0
+_NEON2SSESTORAGE uint32x4_t vrsqrteq_u32(uint32x4_t a); // VRSQRTE.U32 q0,q0
 //Logical operations
 //Bitwise not
-int8x8_t vmvn_s8(int8x8_t a); // VMVN d0,d0
-int16x4_t vmvn_s16(int16x4_t a); // VMVN d0,d0
-int32x2_t vmvn_s32(int32x2_t a); // VMVN d0,d0
-uint8x8_t vmvn_u8(uint8x8_t a); // VMVN d0,d0
-uint16x4_t vmvn_u16(uint16x4_t a); // VMVN d0,d0
-uint32x2_t vmvn_u32(uint32x2_t a); // VMVN d0,d0
-poly8x8_t vmvn_p8(poly8x8_t a); // VMVN d0,d0
-int8x16_t vmvnq_s8(int8x16_t a); // VMVN q0,q0
-int16x8_t vmvnq_s16(int16x8_t a); // VMVN q0,q0
-int32x4_t vmvnq_s32(int32x4_t a); // VMVN q0,q0
-uint8x16_t vmvnq_u8(uint8x16_t a); // VMVN q0,q0
-uint16x8_t vmvnq_u16(uint16x8_t a); // VMVN q0,q0
-uint32x4_t vmvnq_u32(uint32x4_t a); // VMVN q0,q0
-poly8x16_t vmvnq_p8(poly8x16_t a); // VMVN q0,q0
+_NEON2SSESTORAGE int8x8_t vmvn_s8(int8x8_t a); // VMVN d0,d0
+_NEON2SSESTORAGE int16x4_t vmvn_s16(int16x4_t a); // VMVN d0,d0
+_NEON2SSESTORAGE int32x2_t vmvn_s32(int32x2_t a); // VMVN d0,d0
+_NEON2SSESTORAGE uint8x8_t vmvn_u8(uint8x8_t a); // VMVN d0,d0
+_NEON2SSESTORAGE uint16x4_t vmvn_u16(uint16x4_t a); // VMVN d0,d0
+_NEON2SSESTORAGE uint32x2_t vmvn_u32(uint32x2_t a); // VMVN d0,d0
+_NEON2SSESTORAGE poly8x8_t vmvn_p8(poly8x8_t a); // VMVN d0,d0
+_NEON2SSESTORAGE int8x16_t vmvnq_s8(int8x16_t a); // VMVN q0,q0
+_NEON2SSESTORAGE int16x8_t vmvnq_s16(int16x8_t a); // VMVN q0,q0
+_NEON2SSESTORAGE int32x4_t vmvnq_s32(int32x4_t a); // VMVN q0,q0
+_NEON2SSESTORAGE uint8x16_t vmvnq_u8(uint8x16_t a); // VMVN q0,q0
+_NEON2SSESTORAGE uint16x8_t vmvnq_u16(uint16x8_t a); // VMVN q0,q0
+_NEON2SSESTORAGE uint32x4_t vmvnq_u32(uint32x4_t a); // VMVN q0,q0
+_NEON2SSESTORAGE poly8x16_t vmvnq_p8(poly8x16_t a); // VMVN q0,q0
 //Bitwise and
-int8x8_t vand_s8(int8x8_t a, int8x8_t b); // VAND d0,d0,d0
-int16x4_t vand_s16(int16x4_t a, int16x4_t b); // VAND d0,d0,d0
-int32x2_t vand_s32(int32x2_t a, int32x2_t b); // VAND d0,d0,d0
-int64x1_t vand_s64(int64x1_t a, int64x1_t b); // VAND d0,d0,d0
-uint8x8_t vand_u8(uint8x8_t a, uint8x8_t b); // VAND d0,d0,d0
-uint16x4_t vand_u16(uint16x4_t a, uint16x4_t b); // VAND d0,d0,d0
-uint32x2_t vand_u32(uint32x2_t a, uint32x2_t b); // VAND d0,d0,d0
-uint64x1_t vand_u64(uint64x1_t a, uint64x1_t b); // VAND d0,d0,d0
-int8x16_t vandq_s8(int8x16_t a, int8x16_t b); // VAND q0,q0,q0
-int16x8_t vandq_s16(int16x8_t a, int16x8_t b); // VAND q0,q0,q0
-int32x4_t vandq_s32(int32x4_t a, int32x4_t b); // VAND q0,q0,q0
-int64x2_t vandq_s64(int64x2_t a, int64x2_t b); // VAND q0,q0,q0
-uint8x16_t vandq_u8(uint8x16_t a, uint8x16_t b); // VAND q0,q0,q0
-uint16x8_t vandq_u16(uint16x8_t a, uint16x8_t b); // VAND q0,q0,q0
-uint32x4_t vandq_u32(uint32x4_t a, uint32x4_t b); // VAND q0,q0,q0
-uint64x2_t vandq_u64(uint64x2_t a, uint64x2_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vand_s8(int8x8_t a, int8x8_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vand_s16(int16x4_t a, int16x4_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vand_s32(int32x2_t a, int32x2_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vand_s64(int64x1_t a, int64x1_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vand_u8(uint8x8_t a, uint8x8_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vand_u16(uint16x4_t a, uint16x4_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vand_u32(uint32x2_t a, uint32x2_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vand_u64(uint64x1_t a, uint64x1_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vandq_s8(int8x16_t a, int8x16_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vandq_s16(int16x8_t a, int16x8_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vandq_s32(int32x4_t a, int32x4_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vandq_s64(int64x2_t a, int64x2_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vandq_u8(uint8x16_t a, uint8x16_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vandq_u16(uint16x8_t a, uint16x8_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vandq_u32(uint32x4_t a, uint32x4_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vandq_u64(uint64x2_t a, uint64x2_t b); // VAND q0,q0,q0
 //Bitwise or
-int8x8_t vorr_s8(int8x8_t a, int8x8_t b); // VORR d0,d0,d0
-int16x4_t vorr_s16(int16x4_t a, int16x4_t b); // VORR d0,d0,d0
-int32x2_t vorr_s32(int32x2_t a, int32x2_t b); // VORR d0,d0,d0
-int64x1_t vorr_s64(int64x1_t a, int64x1_t b); // VORR d0,d0,d0
-uint8x8_t vorr_u8(uint8x8_t a, uint8x8_t b); // VORR d0,d0,d0
-uint16x4_t vorr_u16(uint16x4_t a, uint16x4_t b); // VORR d0,d0,d0
-uint32x2_t vorr_u32(uint32x2_t a, uint32x2_t b); // VORR d0,d0,d0
-uint64x1_t vorr_u64(uint64x1_t a, uint64x1_t b); // VORR d0,d0,d0
-int8x16_t vorrq_s8(int8x16_t a, int8x16_t b); // VORR q0,q0,q0
-int16x8_t vorrq_s16(int16x8_t a, int16x8_t b); // VORR q0,q0,q0
-int32x4_t vorrq_s32(int32x4_t a, int32x4_t b); // VORR q0,q0,q0
-int64x2_t vorrq_s64(int64x2_t a, int64x2_t b); // VORR q0,q0,q0
-uint8x16_t vorrq_u8(uint8x16_t a, uint8x16_t b); // VORR q0,q0,q0
-uint16x8_t vorrq_u16(uint16x8_t a, uint16x8_t b); // VORR q0,q0,q0
-uint32x4_t vorrq_u32(uint32x4_t a, uint32x4_t b); // VORR q0,q0,q0
-uint64x2_t vorrq_u64(uint64x2_t a, uint64x2_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vorr_s8(int8x8_t a, int8x8_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vorr_s16(int16x4_t a, int16x4_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vorr_s32(int32x2_t a, int32x2_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vorr_s64(int64x1_t a, int64x1_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vorr_u8(uint8x8_t a, uint8x8_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vorr_u16(uint16x4_t a, uint16x4_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vorr_u32(uint32x2_t a, uint32x2_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vorr_u64(uint64x1_t a, uint64x1_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vorrq_s8(int8x16_t a, int8x16_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vorrq_s16(int16x8_t a, int16x8_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vorrq_s32(int32x4_t a, int32x4_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vorrq_s64(int64x2_t a, int64x2_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vorrq_u8(uint8x16_t a, uint8x16_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vorrq_u16(uint16x8_t a, uint16x8_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vorrq_u32(uint32x4_t a, uint32x4_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vorrq_u64(uint64x2_t a, uint64x2_t b); // VORR q0,q0,q0
 //Bitwise exclusive or (EOR or XOR)
-int8x8_t veor_s8(int8x8_t a, int8x8_t b); // VEOR d0,d0,d0
-int16x4_t veor_s16(int16x4_t a, int16x4_t b); // VEOR d0,d0,d0
-int32x2_t veor_s32(int32x2_t a, int32x2_t b); // VEOR d0,d0,d0
-int64x1_t veor_s64(int64x1_t a, int64x1_t b); // VEOR d0,d0,d0
-uint8x8_t veor_u8(uint8x8_t a, uint8x8_t b); // VEOR d0,d0,d0
-uint16x4_t veor_u16(uint16x4_t a, uint16x4_t b); // VEOR d0,d0,d0
-uint32x2_t veor_u32(uint32x2_t a, uint32x2_t b); // VEOR d0,d0,d0
-uint64x1_t veor_u64(uint64x1_t a, uint64x1_t b); // VEOR d0,d0,d0
-int8x16_t veorq_s8(int8x16_t a, int8x16_t b); // VEOR q0,q0,q0
-int16x8_t veorq_s16(int16x8_t a, int16x8_t b); // VEOR q0,q0,q0
-int32x4_t veorq_s32(int32x4_t a, int32x4_t b); // VEOR q0,q0,q0
-int64x2_t veorq_s64(int64x2_t a, int64x2_t b); // VEOR q0,q0,q0
-uint8x16_t veorq_u8(uint8x16_t a, uint8x16_t b); // VEOR q0,q0,q0
-uint16x8_t veorq_u16(uint16x8_t a, uint16x8_t b); // VEOR q0,q0,q0
-uint32x4_t veorq_u32(uint32x4_t a, uint32x4_t b); // VEOR q0,q0,q0
-uint64x2_t veorq_u64(uint64x2_t a, uint64x2_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE int8x8_t veor_s8(int8x8_t a, int8x8_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE int16x4_t veor_s16(int16x4_t a, int16x4_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE int32x2_t veor_s32(int32x2_t a, int32x2_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE int64x1_t veor_s64(int64x1_t a, int64x1_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t veor_u8(uint8x8_t a, uint8x8_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t veor_u16(uint16x4_t a, uint16x4_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t veor_u32(uint32x2_t a, uint32x2_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t veor_u64(uint64x1_t a, uint64x1_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE int8x16_t veorq_s8(int8x16_t a, int8x16_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE int16x8_t veorq_s16(int16x8_t a, int16x8_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE int32x4_t veorq_s32(int32x4_t a, int32x4_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE int64x2_t veorq_s64(int64x2_t a, int64x2_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t veorq_u8(uint8x16_t a, uint8x16_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t veorq_u16(uint16x8_t a, uint16x8_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t veorq_u32(uint32x4_t a, uint32x4_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t veorq_u64(uint64x2_t a, uint64x2_t b); // VEOR q0,q0,q0
 //Bit Clear
-int8x8_t vbic_s8(int8x8_t a, int8x8_t b); // VBIC d0,d0,d0
-int16x4_t vbic_s16(int16x4_t a, int16x4_t b); // VBIC d0,d0,d0
-int32x2_t vbic_s32(int32x2_t a, int32x2_t b); // VBIC d0,d0,d0
-int64x1_t vbic_s64(int64x1_t a, int64x1_t b); // VBIC d0,d0,d0
-uint8x8_t vbic_u8(uint8x8_t a, uint8x8_t b); // VBIC d0,d0,d0
-uint16x4_t vbic_u16(uint16x4_t a, uint16x4_t b); // VBIC d0,d0,d0
-uint32x2_t vbic_u32(uint32x2_t a, uint32x2_t b); // VBIC d0,d0,d0
-uint64x1_t vbic_u64(uint64x1_t a, uint64x1_t b); // VBIC d0,d0,d0
-int8x16_t vbicq_s8(int8x16_t a, int8x16_t b); // VBIC q0,q0,q0
-int16x8_t vbicq_s16(int16x8_t a, int16x8_t b); // VBIC q0,q0,q0
-int32x4_t vbicq_s32(int32x4_t a, int32x4_t b); // VBIC q0,q0,q0
-int64x2_t vbicq_s64(int64x2_t a, int64x2_t b); // VBIC q0,q0,q0
-uint8x16_t vbicq_u8(uint8x16_t a, uint8x16_t b); // VBIC q0,q0,q0
-uint16x8_t vbicq_u16(uint16x8_t a, uint16x8_t b); // VBIC q0,q0,q0
-uint32x4_t vbicq_u32(uint32x4_t a, uint32x4_t b); // VBIC q0,q0,q0
-uint64x2_t vbicq_u64(uint64x2_t a, uint64x2_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vbic_s8(int8x8_t a, int8x8_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vbic_s16(int16x4_t a, int16x4_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vbic_s32(int32x2_t a, int32x2_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vbic_s64(int64x1_t a, int64x1_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vbic_u8(uint8x8_t a, uint8x8_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vbic_u16(uint16x4_t a, uint16x4_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vbic_u32(uint32x2_t a, uint32x2_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vbic_u64(uint64x1_t a, uint64x1_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vbicq_s8(int8x16_t a, int8x16_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vbicq_s16(int16x8_t a, int16x8_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vbicq_s32(int32x4_t a, int32x4_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vbicq_s64(int64x2_t a, int64x2_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vbicq_u8(uint8x16_t a, uint8x16_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vbicq_u16(uint16x8_t a, uint16x8_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vbicq_u32(uint32x4_t a, uint32x4_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vbicq_u64(uint64x2_t a, uint64x2_t b); // VBIC q0,q0,q0
 //Bitwise OR complement
-int8x8_t vorn_s8(int8x8_t a, int8x8_t b); // VORN d0,d0,d0
-int16x4_t vorn_s16(int16x4_t a, int16x4_t b); // VORN d0,d0,d0
-int32x2_t vorn_s32(int32x2_t a, int32x2_t b); // VORN d0,d0,d0
-int64x1_t vorn_s64(int64x1_t a, int64x1_t b); // VORN d0,d0,d0
-uint8x8_t vorn_u8(uint8x8_t a, uint8x8_t b); // VORN d0,d0,d0
-uint16x4_t vorn_u16(uint16x4_t a, uint16x4_t b); // VORN d0,d0,d0
-uint32x2_t vorn_u32(uint32x2_t a, uint32x2_t b); // VORN d0,d0,d0
-uint64x1_t vorn_u64(uint64x1_t a, uint64x1_t b); // VORN d0,d0,d0
-int8x16_t vornq_s8(int8x16_t a, int8x16_t b); // VORN q0,q0,q0
-int16x8_t vornq_s16(int16x8_t a, int16x8_t b); // VORN q0,q0,q0
-int32x4_t vornq_s32(int32x4_t a, int32x4_t b); // VORN q0,q0,q0
-int64x2_t vornq_s64(int64x2_t a, int64x2_t b); // VORN q0,q0,q0
-uint8x16_t vornq_u8(uint8x16_t a, uint8x16_t b); // VORN q0,q0,q0
-uint16x8_t vornq_u16(uint16x8_t a, uint16x8_t b); // VORN q0,q0,q0
-uint32x4_t vornq_u32(uint32x4_t a, uint32x4_t b); // VORN q0,q0,q0
-uint64x2_t vornq_u64(uint64x2_t a, uint64x2_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vorn_s8(int8x8_t a, int8x8_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vorn_s16(int16x4_t a, int16x4_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vorn_s32(int32x2_t a, int32x2_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vorn_s64(int64x1_t a, int64x1_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vorn_u8(uint8x8_t a, uint8x8_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vorn_u16(uint16x4_t a, uint16x4_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vorn_u32(uint32x2_t a, uint32x2_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vorn_u64(uint64x1_t a, uint64x1_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vornq_s8(int8x16_t a, int8x16_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vornq_s16(int16x8_t a, int16x8_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vornq_s32(int32x4_t a, int32x4_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vornq_s64(int64x2_t a, int64x2_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vornq_u8(uint8x16_t a, uint8x16_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vornq_u16(uint16x8_t a, uint16x8_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vornq_u32(uint32x4_t a, uint32x4_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vornq_u64(uint64x2_t a, uint64x2_t b); // VORN q0,q0,q0
 //Bitwise Select
-int8x8_t vbsl_s8(uint8x8_t a, int8x8_t b, int8x8_t c); // VBSL d0,d0,d0
-int16x4_t vbsl_s16(uint16x4_t a, int16x4_t b, int16x4_t c); // VBSL d0,d0,d0
-int32x2_t vbsl_s32(uint32x2_t a, int32x2_t b, int32x2_t c); // VBSL d0,d0,d0
-int64x1_t vbsl_s64(uint64x1_t a, int64x1_t b, int64x1_t c); // VBSL d0,d0,d0
-uint8x8_t vbsl_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VBSL d0,d0,d0
-uint16x4_t vbsl_u16(uint16x4_t a, uint16x4_t b, uint16x4_t c); // VBSL d0,d0,d0
-uint32x2_t vbsl_u32(uint32x2_t a, uint32x2_t b, uint32x2_t c); // VBSL d0,d0,d0
-uint64x1_t vbsl_u64(uint64x1_t a, uint64x1_t b, uint64x1_t c); // VBSL d0,d0,d0
-float32x2_t vbsl_f32(uint32x2_t a, float32x2_t b, float32x2_t c); // VBSL d0,d0,d0
-poly8x8_t vbsl_p8(uint8x8_t a, poly8x8_t b, poly8x8_t c); // VBSL d0,d0,d0
-poly16x4_t vbsl_p16(uint16x4_t a, poly16x4_t b, poly16x4_t c); // VBSL d0,d0,d0
-int8x16_t vbslq_s8(uint8x16_t a, int8x16_t b, int8x16_t c); // VBSL q0,q0,q0
-int16x8_t vbslq_s16(uint16x8_t a, int16x8_t b, int16x8_t c); // VBSL q0,q0,q0
-int32x4_t vbslq_s32(uint32x4_t a, int32x4_t b, int32x4_t c); // VBSL q0,q0,q0
-int64x2_t vbslq_s64(uint64x2_t a, int64x2_t b, int64x2_t c); // VBSL q0,q0,q0
-uint8x16_t vbslq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VBSL q0,q0,q0
-uint16x8_t vbslq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VBSL q0,q0,q0
-uint32x4_t vbslq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VBSL q0,q0,q0
-uint64x2_t vbslq_u64(uint64x2_t a, uint64x2_t b, uint64x2_t c); // VBSL q0,q0,q0
-float32x4_t vbslq_f32(uint32x4_t a, float32x4_t b, float32x4_t c); // VBSL q0,q0,q0
-poly8x16_t vbslq_p8(uint8x16_t a, poly8x16_t b, poly8x16_t c); // VBSL q0,q0,q0
-poly16x8_t vbslq_p16(uint16x8_t a, poly16x8_t b, poly16x8_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE int8x8_t vbsl_s8(uint8x8_t a, int8x8_t b, int8x8_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vbsl_s16(uint16x4_t a, int16x4_t b, int16x4_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vbsl_s32(uint32x2_t a, int32x2_t b, int32x2_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vbsl_s64(uint64x1_t a, int64x1_t b, int64x1_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vbsl_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vbsl_u16(uint16x4_t a, uint16x4_t b, uint16x4_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vbsl_u32(uint32x2_t a, uint32x2_t b, uint32x2_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vbsl_u64(uint64x1_t a, uint64x1_t b, uint64x1_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vbsl_f32(uint32x2_t a, float32x2_t b, float32x2_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE poly8x8_t vbsl_p8(uint8x8_t a, poly8x8_t b, poly8x8_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE poly16x4_t vbsl_p16(uint16x4_t a, poly16x4_t b, poly16x4_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE int8x16_t vbslq_s8(uint8x16_t a, int8x16_t b, int8x16_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vbslq_s16(uint16x8_t a, int16x8_t b, int16x8_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vbslq_s32(uint32x4_t a, int32x4_t b, int32x4_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vbslq_s64(uint64x2_t a, int64x2_t b, int64x2_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vbslq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vbslq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vbslq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vbslq_u64(uint64x2_t a, uint64x2_t b, uint64x2_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vbslq_f32(uint32x4_t a, float32x4_t b, float32x4_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE poly8x16_t vbslq_p8(uint8x16_t a, poly8x16_t b, poly8x16_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE poly16x8_t vbslq_p16(uint16x8_t a, poly16x8_t b, poly16x8_t c); // VBSL q0,q0,q0
 //Transposition operations
 //Transpose elements
-int8x8x2_t vtrn_s8(int8x8_t a, int8x8_t b); // VTRN.8 d0,d0
-int16x4x2_t vtrn_s16(int16x4_t a, int16x4_t b); // VTRN.16 d0,d0
-int32x2x2_t vtrn_s32(int32x2_t a, int32x2_t b); // VTRN.32 d0,d0
-uint8x8x2_t vtrn_u8(uint8x8_t a, uint8x8_t b); // VTRN.8 d0,d0
-uint16x4x2_t vtrn_u16(uint16x4_t a, uint16x4_t b); // VTRN.16 d0,d0
-uint32x2x2_t vtrn_u32(uint32x2_t a, uint32x2_t b); // VTRN.32 d0,d0
-float32x2x2_t vtrn_f32(float32x2_t a, float32x2_t b); // VTRN.32 d0,d0
-poly8x8x2_t vtrn_p8(poly8x8_t a, poly8x8_t b); // VTRN.8 d0,d0
-poly16x4x2_t vtrn_p16(poly16x4_t a, poly16x4_t b); // VTRN.16 d0,d0
-int8x16x2_t vtrnq_s8(int8x16_t a, int8x16_t b); // VTRN.8 q0,q0
-int16x8x2_t vtrnq_s16(int16x8_t a, int16x8_t b); // VTRN.16 q0,q0
-int32x4x2_t vtrnq_s32(int32x4_t a, int32x4_t b); // VTRN.32 q0,q0
-uint8x16x2_t vtrnq_u8(uint8x16_t a, uint8x16_t b); // VTRN.8 q0,q0
-uint16x8x2_t vtrnq_u16(uint16x8_t a, uint16x8_t b); // VTRN.16 q0,q0
-uint32x4x2_t vtrnq_u32(uint32x4_t a, uint32x4_t b); // VTRN.32 q0,q0
-float32x4x2_t vtrnq_f32(float32x4_t a, float32x4_t b); // VTRN.32 q0,q0
-poly8x16x2_t vtrnq_p8(poly8x16_t a, poly8x16_t b); // VTRN.8 q0,q0
-poly16x8x2_t vtrnq_p16(poly16x8_t a, poly16x8_t b); // VTRN.16 q0,q0
+_NEON2SSESTORAGE int8x8x2_t vtrn_s8(int8x8_t a, int8x8_t b); // VTRN.8 d0,d0
+_NEON2SSESTORAGE int16x4x2_t vtrn_s16(int16x4_t a, int16x4_t b); // VTRN.16 d0,d0
+_NEON2SSESTORAGE int32x2x2_t vtrn_s32(int32x2_t a, int32x2_t b); // VTRN.32 d0,d0
+_NEON2SSESTORAGE uint8x8x2_t vtrn_u8(uint8x8_t a, uint8x8_t b); // VTRN.8 d0,d0
+_NEON2SSESTORAGE uint16x4x2_t vtrn_u16(uint16x4_t a, uint16x4_t b); // VTRN.16 d0,d0
+_NEON2SSESTORAGE uint32x2x2_t vtrn_u32(uint32x2_t a, uint32x2_t b); // VTRN.32 d0,d0
+_NEON2SSESTORAGE float32x2x2_t vtrn_f32(float32x2_t a, float32x2_t b); // VTRN.32 d0,d0
+_NEON2SSESTORAGE poly8x8x2_t vtrn_p8(poly8x8_t a, poly8x8_t b); // VTRN.8 d0,d0
+_NEON2SSESTORAGE poly16x4x2_t vtrn_p16(poly16x4_t a, poly16x4_t b); // VTRN.16 d0,d0
+_NEON2SSESTORAGE int8x16x2_t vtrnq_s8(int8x16_t a, int8x16_t b); // VTRN.8 q0,q0
+_NEON2SSESTORAGE int16x8x2_t vtrnq_s16(int16x8_t a, int16x8_t b); // VTRN.16 q0,q0
+_NEON2SSESTORAGE int32x4x2_t vtrnq_s32(int32x4_t a, int32x4_t b); // VTRN.32 q0,q0
+_NEON2SSESTORAGE uint8x16x2_t vtrnq_u8(uint8x16_t a, uint8x16_t b); // VTRN.8 q0,q0
+_NEON2SSESTORAGE uint16x8x2_t vtrnq_u16(uint16x8_t a, uint16x8_t b); // VTRN.16 q0,q0
+_NEON2SSESTORAGE uint32x4x2_t vtrnq_u32(uint32x4_t a, uint32x4_t b); // VTRN.32 q0,q0
+_NEON2SSESTORAGE float32x4x2_t vtrnq_f32(float32x4_t a, float32x4_t b); // VTRN.32 q0,q0
+_NEON2SSESTORAGE poly8x16x2_t vtrnq_p8(poly8x16_t a, poly8x16_t b); // VTRN.8 q0,q0
+_NEON2SSESTORAGE poly16x8x2_t vtrnq_p16(poly16x8_t a, poly16x8_t b); // VTRN.16 q0,q0
 //Interleave elements
-int8x8x2_t vzip_s8(int8x8_t a, int8x8_t b); // VZIP.8 d0,d0
-int16x4x2_t vzip_s16(int16x4_t a, int16x4_t b); // VZIP.16 d0,d0
-int32x2x2_t vzip_s32(int32x2_t a, int32x2_t b); // VZIP.32 d0,d0
-uint8x8x2_t vzip_u8(uint8x8_t a, uint8x8_t b); // VZIP.8 d0,d0
-uint16x4x2_t vzip_u16(uint16x4_t a, uint16x4_t b); // VZIP.16 d0,d0
-uint32x2x2_t vzip_u32(uint32x2_t a, uint32x2_t b); // VZIP.32 d0,d0
-float32x2x2_t vzip_f32(float32x2_t a, float32x2_t b); // VZIP.32 d0,d0
-poly8x8x2_t vzip_p8(poly8x8_t a, poly8x8_t b); // VZIP.8 d0,d0
-poly16x4x2_t vzip_p16(poly16x4_t a, poly16x4_t b); // VZIP.16 d0,d0
-int8x16x2_t vzipq_s8(int8x16_t a, int8x16_t b); // VZIP.8 q0,q0
-int16x8x2_t vzipq_s16(int16x8_t a, int16x8_t b); // VZIP.16 q0,q0
-int32x4x2_t vzipq_s32(int32x4_t a, int32x4_t b); // VZIP.32 q0,q0
-uint8x16x2_t vzipq_u8(uint8x16_t a, uint8x16_t b); // VZIP.8 q0,q0
-uint16x8x2_t vzipq_u16(uint16x8_t a, uint16x8_t b); // VZIP.16 q0,q0
-uint32x4x2_t vzipq_u32(uint32x4_t a, uint32x4_t b); // VZIP.32 q0,q0
-float32x4x2_t vzipq_f32(float32x4_t a, float32x4_t b); // VZIP.32 q0,q0
-poly8x16x2_t vzipq_p8(poly8x16_t a, poly8x16_t b); // VZIP.8 q0,q0
-poly16x8x2_t vzipq_p16(poly16x8_t a, poly16x8_t b); // VZIP.16 q0,q0
+_NEON2SSESTORAGE int8x8x2_t vzip_s8(int8x8_t a, int8x8_t b); // VZIP.8 d0,d0
+_NEON2SSESTORAGE int16x4x2_t vzip_s16(int16x4_t a, int16x4_t b); // VZIP.16 d0,d0
+_NEON2SSESTORAGE int32x2x2_t vzip_s32(int32x2_t a, int32x2_t b); // VZIP.32 d0,d0
+_NEON2SSESTORAGE uint8x8x2_t vzip_u8(uint8x8_t a, uint8x8_t b); // VZIP.8 d0,d0
+_NEON2SSESTORAGE uint16x4x2_t vzip_u16(uint16x4_t a, uint16x4_t b); // VZIP.16 d0,d0
+_NEON2SSESTORAGE uint32x2x2_t vzip_u32(uint32x2_t a, uint32x2_t b); // VZIP.32 d0,d0
+_NEON2SSESTORAGE float32x2x2_t vzip_f32(float32x2_t a, float32x2_t b); // VZIP.32 d0,d0
+_NEON2SSESTORAGE poly8x8x2_t vzip_p8(poly8x8_t a, poly8x8_t b); // VZIP.8 d0,d0
+_NEON2SSESTORAGE poly16x4x2_t vzip_p16(poly16x4_t a, poly16x4_t b); // VZIP.16 d0,d0
+_NEON2SSESTORAGE int8x16x2_t vzipq_s8(int8x16_t a, int8x16_t b); // VZIP.8 q0,q0
+_NEON2SSESTORAGE int16x8x2_t vzipq_s16(int16x8_t a, int16x8_t b); // VZIP.16 q0,q0
+_NEON2SSESTORAGE int32x4x2_t vzipq_s32(int32x4_t a, int32x4_t b); // VZIP.32 q0,q0
+_NEON2SSESTORAGE uint8x16x2_t vzipq_u8(uint8x16_t a, uint8x16_t b); // VZIP.8 q0,q0
+_NEON2SSESTORAGE uint16x8x2_t vzipq_u16(uint16x8_t a, uint16x8_t b); // VZIP.16 q0,q0
+_NEON2SSESTORAGE uint32x4x2_t vzipq_u32(uint32x4_t a, uint32x4_t b); // VZIP.32 q0,q0
+_NEON2SSESTORAGE float32x4x2_t vzipq_f32(float32x4_t a, float32x4_t b); // VZIP.32 q0,q0
+_NEON2SSESTORAGE poly8x16x2_t vzipq_p8(poly8x16_t a, poly8x16_t b); // VZIP.8 q0,q0
+_NEON2SSESTORAGE poly16x8x2_t vzipq_p16(poly16x8_t a, poly16x8_t b); // VZIP.16 q0,q0
 //De-Interleave elements
-int8x8x2_t vuzp_s8(int8x8_t a, int8x8_t b); // VUZP.8 d0,d0
-int16x4x2_t vuzp_s16(int16x4_t a, int16x4_t b); // VUZP.16 d0,d0
-int32x2x2_t vuzp_s32(int32x2_t a, int32x2_t b); // VUZP.32 d0,d0
-uint8x8x2_t vuzp_u8(uint8x8_t a, uint8x8_t b); // VUZP.8 d0,d0
-uint16x4x2_t vuzp_u16(uint16x4_t a, uint16x4_t b); // VUZP.16 d0,d0
-uint32x2x2_t vuzp_u32(uint32x2_t a, uint32x2_t b); // VUZP.32 d0,d0
-float32x2x2_t vuzp_f32(float32x2_t a, float32x2_t b); // VUZP.32 d0,d0
-poly8x8x2_t vuzp_p8(poly8x8_t a, poly8x8_t b); // VUZP.8 d0,d0
-poly16x4x2_t vuzp_p16(poly16x4_t a, poly16x4_t b); // VUZP.16 d0,d0
-int8x16x2_t vuzpq_s8(int8x16_t a, int8x16_t b); // VUZP.8 q0,q0
-int16x8x2_t vuzpq_s16(int16x8_t a, int16x8_t b); // VUZP.16 q0,q0
-int32x4x2_t vuzpq_s32(int32x4_t a, int32x4_t b); // VUZP.32 q0,q0
-uint8x16x2_t vuzpq_u8(uint8x16_t a, uint8x16_t b); // VUZP.8 q0,q0
-uint16x8x2_t vuzpq_u16(uint16x8_t a, uint16x8_t b); // VUZP.16 q0,q0
-uint32x4x2_t vuzpq_u32(uint32x4_t a, uint32x4_t b); // VUZP.32 q0,q0
-float32x4x2_t vuzpq_f32(float32x4_t a, float32x4_t b); // VUZP.32 q0,q0
-poly8x16x2_t vuzpq_p8(poly8x16_t a, poly8x16_t b); // VUZP.8 q0,q0
-poly16x8x2_t vuzpq_p16(poly16x8_t a, poly16x8_t b); // VUZP.16 q0,q0
-
-float32x4_t vrndnq_f32(float32x4_t a); // VRND.F32 q0,q0
-
-float64x2_t vrndnq_f64(float64x2_t a); // VRND.F64 q0,q0
+_NEON2SSESTORAGE int8x8x2_t vuzp_s8(int8x8_t a, int8x8_t b); // VUZP.8 d0,d0
+_NEON2SSESTORAGE int16x4x2_t vuzp_s16(int16x4_t a, int16x4_t b); // VUZP.16 d0,d0
+_NEON2SSESTORAGE int32x2x2_t vuzp_s32(int32x2_t a, int32x2_t b); // VUZP.32 d0,d0
+_NEON2SSESTORAGE uint8x8x2_t vuzp_u8(uint8x8_t a, uint8x8_t b); // VUZP.8 d0,d0
+_NEON2SSESTORAGE uint16x4x2_t vuzp_u16(uint16x4_t a, uint16x4_t b); // VUZP.16 d0,d0
+_NEON2SSESTORAGE uint32x2x2_t vuzp_u32(uint32x2_t a, uint32x2_t b); // VUZP.32 d0,d0
+_NEON2SSESTORAGE float32x2x2_t vuzp_f32(float32x2_t a, float32x2_t b); // VUZP.32 d0,d0
+_NEON2SSESTORAGE poly8x8x2_t vuzp_p8(poly8x8_t a, poly8x8_t b); // VUZP.8 d0,d0
+_NEON2SSESTORAGE poly16x4x2_t vuzp_p16(poly16x4_t a, poly16x4_t b); // VUZP.16 d0,d0
+_NEON2SSESTORAGE int8x16x2_t vuzpq_s8(int8x16_t a, int8x16_t b); // VUZP.8 q0,q0
+_NEON2SSESTORAGE int16x8x2_t vuzpq_s16(int16x8_t a, int16x8_t b); // VUZP.16 q0,q0
+_NEON2SSESTORAGE int32x4x2_t vuzpq_s32(int32x4_t a, int32x4_t b); // VUZP.32 q0,q0
+_NEON2SSESTORAGE uint8x16x2_t vuzpq_u8(uint8x16_t a, uint8x16_t b); // VUZP.8 q0,q0
+_NEON2SSESTORAGE uint16x8x2_t vuzpq_u16(uint16x8_t a, uint16x8_t b); // VUZP.16 q0,q0
+_NEON2SSESTORAGE uint32x4x2_t vuzpq_u32(uint32x4_t a, uint32x4_t b); // VUZP.32 q0,q0
+_NEON2SSESTORAGE float32x4x2_t vuzpq_f32(float32x4_t a, float32x4_t b); // VUZP.32 q0,q0
+_NEON2SSESTORAGE poly8x16x2_t vuzpq_p8(poly8x16_t a, poly8x16_t b); // VUZP.8 q0,q0
+_NEON2SSESTORAGE poly16x8x2_t vuzpq_p16(poly16x8_t a, poly16x8_t b); // VUZP.16 q0,q0
+
+_NEON2SSESTORAGE float32x4_t vrndnq_f32(float32x4_t a); // VRND.F32 q0,q0
+
+_NEON2SSESTORAGE float64x2_t vrndnq_f64(float64x2_t a); // VRND.F64 q0,q0
 
 //Sqrt
-float32x4_t vsqrtq_f32(float32x4_t a); // VSQRT.F32 q0,q0
+_NEON2SSESTORAGE float32x4_t vsqrtq_f32(float32x4_t a); // VSQRT.F32 q0,q0
 
-float64x2_t vsqrtq_f64(float64x2_t a); // VSQRT.F64 q0,q0
+_NEON2SSESTORAGE float64x2_t vsqrtq_f64(float64x2_t a); // VSQRT.F64 q0,q0
 
 
 
 //^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-// the following macros solve the problem of the "immediate parameters requirement" for some x86 intrinsics. 
+// the following macros solve the problem of the "immediate parameters requirement" for some x86 intrinsics.
 // we need it to compile the code unless the "Intrinsic parameter must be an immediate value" error is our goal
 //
-#if  ( defined (__INTEL_COMPILER)  || defined (__GNUC__) && !defined(__llvm__) )    
-
-    #define _MM_ALIGNR_EPI8 _mm_alignr_epi8
-
-#define _MM_EXTRACT_EPI16  (int16_t) _mm_extract_epi16
-    #define _MM_INSERT_EPI16 _mm_insert_epi16
-#ifdef USE_SSE4
-        #define _MM_EXTRACT_EPI8  _mm_extract_epi8
-        #define _MM_EXTRACT_EPI32  _mm_extract_epi32
-        #define _MM_EXTRACT_PS  _mm_extract_ps
-
-        #define _MM_INSERT_EPI8  _mm_insert_epi8
-        #define _MM_INSERT_EPI32 _mm_insert_epi32
-        #define _MM_INSERT_PS    _mm_insert_ps
-#ifdef  _NEON2SSE_64BIT
-            #define _MM_INSERT_EPI64 _mm_insert_epi64
-            #define _MM_EXTRACT_EPI64 _mm_extract_epi64
-#endif
-#endif //SSE4
+#if  ( defined (__INTEL_COMPILER)  || defined (__GNUC__) && !defined(__llvm__) )
+#   define _MM_ALIGNR_EPI8 _mm_alignr_epi8
+#   define _MM_EXTRACT_EPI16  (int16_t) _mm_extract_epi16
+#   define _MM_INSERT_EPI16 _mm_insert_epi16
+#   ifdef USE_SSE4
+#       define _MM_EXTRACT_EPI8  _mm_extract_epi8
+#       define _MM_EXTRACT_EPI32  _mm_extract_epi32
+#       define _MM_EXTRACT_PS  _mm_extract_ps
+#       define _MM_INSERT_EPI8  _mm_insert_epi8
+#       define _MM_INSERT_EPI32 _mm_insert_epi32
+#       define _MM_INSERT_PS    _mm_insert_ps
+#       ifdef  _NEON2SSE_64BIT
+#           define _MM_INSERT_EPI64 _mm_insert_epi64
+#           define _MM_EXTRACT_EPI64 _mm_extract_epi64
+#       endif
+#   endif //SSE4
 #else
-    #define _NEON2SSE_COMMA ,
-    #define _NEON2SSE_SWITCH16(NAME, a, b, LANE) \
-            switch(LANE)         \
+#   define _NEON2SSE_COMMA ,
+#   define _NEON2SSE_SWITCH16(NAME, a, b, LANE) \
+        switch(LANE)         \
         {                \
         case 0:     return NAME(a b, 0); \
         case 1:     return NAME(a b, 1); \
@@ -2325,8 +2327,8 @@ float64x2_t vsqrtq_f64(float64x2_t a); // VSQRT.F64 q0,q0
         default:    return NAME(a b, 0); \
         }
 
-    #define _NEON2SSE_SWITCH8(NAME, vec, LANE, p) \
-            switch(LANE)              \
+#   define _NEON2SSE_SWITCH8(NAME, vec, LANE, p) \
+        switch(LANE)              \
         {                          \
         case 0:  return NAME(vec p,0); \
         case 1:  return NAME(vec p,1); \
@@ -2339,8 +2341,8 @@ float64x2_t vsqrtq_f64(float64x2_t a); // VSQRT.F64 q0,q0
         default: return NAME(vec p,0); \
         }
 
-    #define _NEON2SSE_SWITCH4(NAME, case0, case1, case2, case3, vec, LANE, p) \
-            switch(LANE)              \
+#   define _NEON2SSE_SWITCH4(NAME, case0, case1, case2, case3, vec, LANE, p) \
+        switch(LANE)              \
         {                          \
         case case0:  return NAME(vec p,case0); \
         case case1:  return NAME(vec p,case1); \
@@ -2424,32 +2426,32 @@ float64x2_t vsqrtq_f64(float64x2_t a); // VSQRT.F64 q0,q0
 // Below are some helper functions used either for SSE4 intrinsics "emulation" for SSSE3 limited devices
 // or for some specific commonly used operations implementation missing in SSE
 #ifdef USE_SSE4
-    #define _MM_CVTEPU8_EPI16  _mm_cvtepu8_epi16
-    #define _MM_CVTEPU16_EPI32 _mm_cvtepu16_epi32
-    #define _MM_CVTEPU32_EPI64  _mm_cvtepu32_epi64
+#   define _MM_CVTEPU8_EPI16  _mm_cvtepu8_epi16
+#   define _MM_CVTEPU16_EPI32 _mm_cvtepu16_epi32
+#   define _MM_CVTEPU32_EPI64  _mm_cvtepu32_epi64
 
-    #define _MM_CVTEPI8_EPI16  _mm_cvtepi8_epi16
-    #define _MM_CVTEPI16_EPI32 _mm_cvtepi16_epi32
-    #define _MM_CVTEPI32_EPI64  _mm_cvtepi32_epi64
+#   define _MM_CVTEPI8_EPI16  _mm_cvtepi8_epi16
+#   define _MM_CVTEPI16_EPI32 _mm_cvtepi16_epi32
+#   define _MM_CVTEPI32_EPI64  _mm_cvtepi32_epi64
 
-    #define _MM_MAX_EPI8  _mm_max_epi8
-    #define _MM_MAX_EPI32 _mm_max_epi32
-    #define _MM_MAX_EPU16 _mm_max_epu16
-    #define _MM_MAX_EPU32 _mm_max_epu32
+#   define _MM_MAX_EPI8  _mm_max_epi8
+#   define _MM_MAX_EPI32 _mm_max_epi32
+#   define _MM_MAX_EPU16 _mm_max_epu16
+#   define _MM_MAX_EPU32 _mm_max_epu32
 
-    #define _MM_MIN_EPI8  _mm_min_epi8
-    #define _MM_MIN_EPI32 _mm_min_epi32
-    #define _MM_MIN_EPU16 _mm_min_epu16
-    #define _MM_MIN_EPU32 _mm_min_epu32
+#   define _MM_MIN_EPI8  _mm_min_epi8
+#   define _MM_MIN_EPI32 _mm_min_epi32
+#   define _MM_MIN_EPU16 _mm_min_epu16
+#   define _MM_MIN_EPU32 _mm_min_epu32
 
-    #define _MM_BLENDV_EPI8 _mm_blendv_epi8
-    #define _MM_PACKUS_EPI32 _mm_packus_epi32
-    #define _MM_PACKUS1_EPI32(a) _mm_packus_epi32(a, a)
+#   define _MM_BLENDV_EPI8 _mm_blendv_epi8
+#   define _MM_PACKUS_EPI32 _mm_packus_epi32
+#   define _MM_PACKUS1_EPI32(a) _mm_packus_epi32(a, a)
 
-    #define _MM_MULLO_EPI32 _mm_mullo_epi32
-    #define _MM_MUL_EPI32  _mm_mul_epi32
+#   define _MM_MULLO_EPI32 _mm_mullo_epi32
+#   define _MM_MUL_EPI32  _mm_mul_epi32
 
-    #define _MM_CMPEQ_EPI64 _mm_cmpeq_epi64
+#   define _MM_CMPEQ_EPI64 _mm_cmpeq_epi64
 #else     //no SSE4 !!!!!!
     _NEON2SSE_INLINE __m128i _MM_CVTEPU8_EPI16(__m128i a)
     {
@@ -2729,11 +2731,11 @@ _NEON2SSE_INLINE int64_t _MM_EXTRACT_EPI64_32(__m128i val, const int LANE)
 }
 
 #ifndef _NEON2SSE_64BIT_SSE4
-    #define _MM_INSERT_EPI64 _MM_INSERT_EPI64_32
-    #define _MM_EXTRACT_EPI64 _MM_EXTRACT_EPI64_32
+#   define _MM_INSERT_EPI64 _MM_INSERT_EPI64_32
+#   define _MM_EXTRACT_EPI64 _MM_EXTRACT_EPI64_32
 #endif
 
-int32x4_t  vqd_s32(int32x4_t a); //Doubling saturation for signed ints
+_NEON2SSESTORAGE int32x4_t  vqd_s32(int32x4_t a); //Doubling saturation for signed ints
 _NEON2SSE_INLINE int32x4_t  vqd_s32(int32x4_t a)
 {
     //Overflow happens only if a and sum have the opposite signs
@@ -2780,7 +2782,7 @@ the serial implementation is provided along with the corresponding compiler warn
 //***********************************************************************
 //************************      Vector add   *****************************
 //***********************************************************************
-int8x8_t vadd_s8(int8x8_t a, int8x8_t b); // VADD.I8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vadd_s8(int8x8_t a, int8x8_t b); // VADD.I8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vadd_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
@@ -2788,7 +2790,7 @@ _NEON2SSE_INLINE int8x8_t vadd_s8(int8x8_t a, int8x8_t b)
 }
 
 
-int16x4_t vadd_s16(int16x4_t a, int16x4_t b); // VADD.I16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vadd_s16(int16x4_t a, int16x4_t b); // VADD.I16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vadd_s16(int16x4_t a, int16x4_t b)
 {
     int16x4_t res64;
@@ -2796,7 +2798,7 @@ _NEON2SSE_INLINE int16x4_t vadd_s16(int16x4_t a, int16x4_t b)
 }
 
 
-int32x2_t vadd_s32(int32x2_t a, int32x2_t b); // VADD.I32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vadd_s32(int32x2_t a, int32x2_t b); // VADD.I32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vadd_s32(int32x2_t a, int32x2_t b)
 {
     int32x2_t res64;
@@ -2804,7 +2806,7 @@ _NEON2SSE_INLINE int32x2_t vadd_s32(int32x2_t a, int32x2_t b)
 }
 
 
-int64x1_t  vadd_s64(int64x1_t a,  int64x1_t b); // VADD.I64 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t  vadd_s64(int64x1_t a,  int64x1_t b); // VADD.I64 d0,d0,d0
 _NEON2SSE_INLINE int64x1_t  vadd_s64(int64x1_t a,  int64x1_t b)
 {
     int64x1_t res64;
@@ -2813,7 +2815,7 @@ _NEON2SSE_INLINE int64x1_t  vadd_s64(int64x1_t a,  int64x1_t b)
 }
 
 
-float32x2_t vadd_f32(float32x2_t a, float32x2_t b); // VADD.F32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vadd_f32(float32x2_t a, float32x2_t b); // VADD.F32 d0,d0,d0
 _NEON2SSE_INLINE float32x2_t vadd_f32(float32x2_t a, float32x2_t b)
 {
     __m128 res;
@@ -2823,16 +2825,16 @@ _NEON2SSE_INLINE float32x2_t vadd_f32(float32x2_t a, float32x2_t b)
     return res64;
 }
 
-uint8x8_t  vadd_u8(uint8x8_t a, uint8x8_t b); // VADD.I8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t  vadd_u8(uint8x8_t a, uint8x8_t b); // VADD.I8 d0,d0,d0
 #define vadd_u8 vadd_s8
 
-uint16x4_t  vadd_u16(uint16x4_t a, uint16x4_t b); // VADD.I16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t  vadd_u16(uint16x4_t a, uint16x4_t b); // VADD.I16 d0,d0,d0
 #define vadd_u16 vadd_s16
 
-uint32x2_t  vadd_u32(uint32x2_t a, uint32x2_t b); // VADD.I32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t  vadd_u32(uint32x2_t a, uint32x2_t b); // VADD.I32 d0,d0,d0
 #define vadd_u32 vadd_s32
 
-uint64x1_t vadd_u64(uint64x1_t a,  uint64x1_t b); // VADD.I64 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vadd_u64(uint64x1_t a,  uint64x1_t b); // VADD.I64 d0,d0,d0
 _NEON2SSE_INLINE uint64x1_t vadd_u64(uint64x1_t a,  uint64x1_t b)
 {
     uint64x1_t res64;
@@ -2841,37 +2843,37 @@ _NEON2SSE_INLINE uint64x1_t vadd_u64(uint64x1_t a,  uint64x1_t b)
 }
 
 
-int8x16_t   vaddq_s8(int8x16_t a, int8x16_t b); // VADD.I8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t   vaddq_s8(int8x16_t a, int8x16_t b); // VADD.I8 q0,q0,q0
 #define vaddq_s8 _mm_add_epi8
 
-int16x8_t   vaddq_s16(int16x8_t a, int16x8_t b); // VADD.I16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t   vaddq_s16(int16x8_t a, int16x8_t b); // VADD.I16 q0,q0,q0
 #define vaddq_s16 _mm_add_epi16
 
-int32x4_t   vaddq_s32(int32x4_t a, int32x4_t b); // VADD.I32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t   vaddq_s32(int32x4_t a, int32x4_t b); // VADD.I32 q0,q0,q0
 #define vaddq_s32 _mm_add_epi32
 
-int64x2_t   vaddq_s64(int64x2_t a, int64x2_t b); // VADD.I64 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t   vaddq_s64(int64x2_t a, int64x2_t b); // VADD.I64 q0,q0,q0
 #define vaddq_s64 _mm_add_epi64
 
-float32x4_t vaddq_f32(float32x4_t a, float32x4_t b); // VADD.F32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vaddq_f32(float32x4_t a, float32x4_t b); // VADD.F32 q0,q0,q0
 #define vaddq_f32 _mm_add_ps
 
-uint8x16_t   vaddq_u8(uint8x16_t a, uint8x16_t b); // VADD.I8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t   vaddq_u8(uint8x16_t a, uint8x16_t b); // VADD.I8 q0,q0,q0
 #define vaddq_u8 _mm_add_epi8
 
-uint16x8_t   vaddq_u16(uint16x8_t a, uint16x8_t b); // VADD.I16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   vaddq_u16(uint16x8_t a, uint16x8_t b); // VADD.I16 q0,q0,q0
 #define vaddq_u16 _mm_add_epi16
 
-uint32x4_t   vaddq_u32(uint32x4_t a, uint32x4_t b); // VADD.I32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t   vaddq_u32(uint32x4_t a, uint32x4_t b); // VADD.I32 q0,q0,q0
 #define vaddq_u32 _mm_add_epi32
 
-uint64x2_t   vaddq_u64(uint64x2_t a, uint64x2_t b); // VADD.I64 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t   vaddq_u64(uint64x2_t a, uint64x2_t b); // VADD.I64 q0,q0,q0
 #define vaddq_u64 _mm_add_epi64
 
 //**************************** Vector long add *****************************:
 //***********************************************************************
 //Va, Vb have equal lane sizes, result is a 128 bit vector of lanes that are twice the width.
-int16x8_t  vaddl_s8(int8x8_t a, int8x8_t b); // VADDL.S8 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t  vaddl_s8(int8x8_t a, int8x8_t b); // VADDL.S8 q0,d0,d0
 _NEON2SSE_INLINE int16x8_t  vaddl_s8(int8x8_t a, int8x8_t b) // VADDL.S8 q0,d0,d0
 {
     __m128i a16, b16;
@@ -2880,7 +2882,7 @@ _NEON2SSE_INLINE int16x8_t  vaddl_s8(int8x8_t a, int8x8_t b) // VADDL.S8 q0,d0,d
     return _mm_add_epi16 (a16, b16);
 }
 
-int32x4_t  vaddl_s16(int16x4_t a, int16x4_t b); // VADDL.S16 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t  vaddl_s16(int16x4_t a, int16x4_t b); // VADDL.S16 q0,d0,d0
 _NEON2SSE_INLINE int32x4_t  vaddl_s16(int16x4_t a, int16x4_t b) // VADDL.S16 q0,d0,d0
 {
     __m128i a32, b32;
@@ -2889,7 +2891,7 @@ _NEON2SSE_INLINE int32x4_t  vaddl_s16(int16x4_t a, int16x4_t b) // VADDL.S16 q0,
     return _mm_add_epi32 (a32, b32);
 }
 
-int64x2_t  vaddl_s32(int32x2_t a, int32x2_t b); // VADDL.S32 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t  vaddl_s32(int32x2_t a, int32x2_t b); // VADDL.S32 q0,d0,d0
 _NEON2SSE_INLINE int64x2_t  vaddl_s32(int32x2_t a, int32x2_t b) // VADDL.S32 q0,d0,d0
 {
     //may be not optimal
@@ -2899,7 +2901,7 @@ _NEON2SSE_INLINE int64x2_t  vaddl_s32(int32x2_t a, int32x2_t b) // VADDL.S32 q0,
     return _mm_add_epi64 ( a64, b64);
 }
 
-uint16x8_t vaddl_u8(uint8x8_t a, uint8x8_t b); // VADDL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vaddl_u8(uint8x8_t a, uint8x8_t b); // VADDL.U8 q0,d0,d0
 _NEON2SSE_INLINE uint16x8_t vaddl_u8(uint8x8_t a, uint8x8_t b) // VADDL.U8 q0,d0,d0
 {
     __m128i a16, b16;
@@ -2908,7 +2910,7 @@ _NEON2SSE_INLINE uint16x8_t vaddl_u8(uint8x8_t a, uint8x8_t b) // VADDL.U8 q0,d0
     return _mm_add_epi16 (a16, b16);
 }
 
-uint32x4_t vaddl_u16(uint16x4_t a, uint16x4_t b); // VADDL.s16 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vaddl_u16(uint16x4_t a, uint16x4_t b); // VADDL.s16 q0,d0,d0
 _NEON2SSE_INLINE uint32x4_t vaddl_u16(uint16x4_t a, uint16x4_t b) // VADDL.s16 q0,d0,d0
 {
     __m128i a32, b32;
@@ -2917,7 +2919,7 @@ _NEON2SSE_INLINE uint32x4_t vaddl_u16(uint16x4_t a, uint16x4_t b) // VADDL.s16 q
     return _mm_add_epi32 (a32, b32);
 }
 
-uint64x2_t vaddl_u32(uint32x2_t a, uint32x2_t b); // VADDL.U32 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vaddl_u32(uint32x2_t a, uint32x2_t b); // VADDL.U32 q0,d0,d0
 _NEON2SSE_INLINE uint64x2_t vaddl_u32(uint32x2_t a, uint32x2_t b) // VADDL.U32 q0,d0,d0
 {
     //may be not optimal
@@ -2929,7 +2931,7 @@ _NEON2SSE_INLINE uint64x2_t vaddl_u32(uint32x2_t a, uint32x2_t b) // VADDL.U32 q
 
 //***************   Vector wide add: vaddw_<type>. Vr[i]:=Va[i]+Vb[i] ******************
 //*************** *********************************************************************
-int16x8_t  vaddw_s8(int16x8_t a, int8x8_t b); // VADDW.S8 q0,q0,d0
+_NEON2SSESTORAGE int16x8_t  vaddw_s8(int16x8_t a, int8x8_t b); // VADDW.S8 q0,q0,d0
 _NEON2SSE_INLINE int16x8_t  vaddw_s8(int16x8_t a, int8x8_t b) // VADDW.S8 q0,q0,d0
 {
     __m128i b16;
@@ -2937,7 +2939,7 @@ _NEON2SSE_INLINE int16x8_t  vaddw_s8(int16x8_t a, int8x8_t b) // VADDW.S8 q0,q0,
     return _mm_add_epi16 (a, b16);
 }
 
-int32x4_t  vaddw_s16(int32x4_t a, int16x4_t b); // VADDW.S16 q0,q0,d0
+_NEON2SSESTORAGE int32x4_t  vaddw_s16(int32x4_t a, int16x4_t b); // VADDW.S16 q0,q0,d0
 _NEON2SSE_INLINE int32x4_t  vaddw_s16(int32x4_t a, int16x4_t b) // VADDW.S16 q0,q0,d0
 {
     __m128i b32;
@@ -2945,7 +2947,7 @@ _NEON2SSE_INLINE int32x4_t  vaddw_s16(int32x4_t a, int16x4_t b) // VADDW.S16 q0,
     return _mm_add_epi32 (a, b32);
 }
 
-int64x2_t  vaddw_s32(int64x2_t a, int32x2_t b); // VADDW.S32 q0,q0,d0
+_NEON2SSESTORAGE int64x2_t  vaddw_s32(int64x2_t a, int32x2_t b); // VADDW.S32 q0,q0,d0
 _NEON2SSE_INLINE int64x2_t  vaddw_s32(int64x2_t a, int32x2_t b) // VADDW.S32 q0,q0,d0
 {
     __m128i b64;
@@ -2953,7 +2955,7 @@ _NEON2SSE_INLINE int64x2_t  vaddw_s32(int64x2_t a, int32x2_t b) // VADDW.S32 q0,
     return _mm_add_epi64 (a, b64);
 }
 
-uint16x8_t vaddw_u8(uint16x8_t a, uint8x8_t b); // VADDW.U8 q0,q0,d0
+_NEON2SSESTORAGE uint16x8_t vaddw_u8(uint16x8_t a, uint8x8_t b); // VADDW.U8 q0,q0,d0
 _NEON2SSE_INLINE uint16x8_t vaddw_u8(uint16x8_t a, uint8x8_t b) // VADDW.U8 q0,q0,d0
 {
     __m128i b16;
@@ -2961,7 +2963,7 @@ _NEON2SSE_INLINE uint16x8_t vaddw_u8(uint16x8_t a, uint8x8_t b) // VADDW.U8 q0,q
     return _mm_add_epi16 (a, b16);
 }
 
-uint32x4_t vaddw_u16(uint32x4_t a, uint16x4_t b); // VADDW.s16 q0,q0,d0
+_NEON2SSESTORAGE uint32x4_t vaddw_u16(uint32x4_t a, uint16x4_t b); // VADDW.s16 q0,q0,d0
 _NEON2SSE_INLINE uint32x4_t vaddw_u16(uint32x4_t a, uint16x4_t b) // VADDW.s16 q0,q0,d0
 {
     __m128i b32;
@@ -2969,7 +2971,7 @@ _NEON2SSE_INLINE uint32x4_t vaddw_u16(uint32x4_t a, uint16x4_t b) // VADDW.s16 q
     return _mm_add_epi32 (a, b32);
 }
 
-uint64x2_t vaddw_u32(uint64x2_t a, uint32x2_t b); // VADDW.U32 q0,q0,d0
+_NEON2SSESTORAGE uint64x2_t vaddw_u32(uint64x2_t a, uint32x2_t b); // VADDW.U32 q0,q0,d0
 _NEON2SSE_INLINE uint64x2_t vaddw_u32(uint64x2_t a, uint32x2_t b) // VADDW.U32 q0,q0,d0
 {
     __m128i b64;
@@ -2979,7 +2981,7 @@ _NEON2SSE_INLINE uint64x2_t vaddw_u32(uint64x2_t a, uint32x2_t b) // VADDW.U32 q
 
 //******************************Vector halving add: vhadd -> Vr[i]:=(Va[i]+Vb[i])>>1 ,  result truncated *******************************
 //*************************************************************************************************************************
-int8x8_t vhadd_s8(int8x8_t a,  int8x8_t b); // VHADD.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vhadd_s8(int8x8_t a,  int8x8_t b); // VHADD.S8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vhadd_s8(int8x8_t a,  int8x8_t b)
 {
     int8x8_t res64;
@@ -2987,7 +2989,7 @@ _NEON2SSE_INLINE int8x8_t vhadd_s8(int8x8_t a,  int8x8_t b)
 }
 
 
-int16x4_t vhadd_s16(int16x4_t a,  int16x4_t b); // VHADD.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vhadd_s16(int16x4_t a,  int16x4_t b); // VHADD.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vhadd_s16(int16x4_t a,  int16x4_t b)
 {
     int16x4_t res64;
@@ -2995,7 +2997,7 @@ _NEON2SSE_INLINE int16x4_t vhadd_s16(int16x4_t a,  int16x4_t b)
 }
 
 
-int32x2_t vhadd_s32(int32x2_t a,  int32x2_t b); // VHADD.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vhadd_s32(int32x2_t a,  int32x2_t b); // VHADD.S32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vhadd_s32(int32x2_t a,  int32x2_t b)
 {
     int32x2_t res64;
@@ -3003,7 +3005,7 @@ _NEON2SSE_INLINE int32x2_t vhadd_s32(int32x2_t a,  int32x2_t b)
 }
 
 
-uint8x8_t vhadd_u8(uint8x8_t a,  uint8x8_t b); // VHADD.w d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vhadd_u8(uint8x8_t a,  uint8x8_t b); // VHADD.w d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vhadd_u8(uint8x8_t a,  uint8x8_t b)
 {
     uint8x8_t res64;
@@ -3011,7 +3013,7 @@ _NEON2SSE_INLINE uint8x8_t vhadd_u8(uint8x8_t a,  uint8x8_t b)
 }
 
 
-uint16x4_t vhadd_u16(uint16x4_t a,  uint16x4_t b); // VHADD.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vhadd_u16(uint16x4_t a,  uint16x4_t b); // VHADD.s16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vhadd_u16(uint16x4_t a,  uint16x4_t b)
 {
     uint16x4_t res64;
@@ -3019,7 +3021,7 @@ _NEON2SSE_INLINE uint16x4_t vhadd_u16(uint16x4_t a,  uint16x4_t b)
 }
 
 
-uint32x2_t vhadd_u32(uint32x2_t a,  uint32x2_t b); // VHADD.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vhadd_u32(uint32x2_t a,  uint32x2_t b); // VHADD.U32 d0,d0,d0
 _NEON2SSE_INLINE uint32x2_t vhadd_u32(uint32x2_t a,  uint32x2_t b)
 {
     uint32x2_t res64;
@@ -3027,7 +3029,7 @@ _NEON2SSE_INLINE uint32x2_t vhadd_u32(uint32x2_t a,  uint32x2_t b)
 }
 
 
-int8x16_t vhaddq_s8(int8x16_t a, int8x16_t b); // VHADD.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vhaddq_s8(int8x16_t a, int8x16_t b); // VHADD.S8 q0,q0,q0
 _NEON2SSE_INLINE int8x16_t vhaddq_s8(int8x16_t a, int8x16_t b)
 {
     //need to avoid internal overflow, will use the (x&y)+((x^y)>>1).
@@ -3038,7 +3040,7 @@ _NEON2SSE_INLINE int8x16_t vhaddq_s8(int8x16_t a, int8x16_t b)
     return _mm_add_epi8(tmp1,tmp2);
 }
 
-int16x8_t vhaddq_s16(int16x8_t a, int16x8_t b); // VHADD.S1 6 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vhaddq_s16(int16x8_t a, int16x8_t b); // VHADD.S1 6 q0,q0,q0
 _NEON2SSE_INLINE int16x8_t vhaddq_s16(int16x8_t a, int16x8_t b)
 {
     //need to avoid internal overflow, will use the (x&y)+((x^y)>>1).
@@ -3049,7 +3051,7 @@ _NEON2SSE_INLINE int16x8_t vhaddq_s16(int16x8_t a, int16x8_t b)
     return _mm_add_epi16(tmp1,tmp2);
 }
 
-int32x4_t vhaddq_s32(int32x4_t a, int32x4_t b); // VHADD.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vhaddq_s32(int32x4_t a, int32x4_t b); // VHADD.S32 q0,q0,q0
 _NEON2SSE_INLINE int32x4_t vhaddq_s32(int32x4_t a, int32x4_t b) // VHADD.S32 q0,q0,q0
 {
     //need to avoid internal overflow, will use the (x&y)+((x^y)>>1).
@@ -3060,7 +3062,7 @@ _NEON2SSE_INLINE int32x4_t vhaddq_s32(int32x4_t a, int32x4_t b) // VHADD.S32 q0,
     return _mm_add_epi32(tmp1,tmp2);
 }
 
-uint8x16_t vhaddq_u8(uint8x16_t a, uint8x16_t b); // VHADD.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vhaddq_u8(uint8x16_t a, uint8x16_t b); // VHADD.U8 q0,q0,q0
 _NEON2SSE_INLINE uint8x16_t vhaddq_u8(uint8x16_t a, uint8x16_t b) // VHADD.U8 q0,q0,q0
 {
     __m128i c1, sum, res;
@@ -3071,7 +3073,7 @@ _NEON2SSE_INLINE uint8x16_t vhaddq_u8(uint8x16_t a, uint8x16_t b) // VHADD.U8 q0
     return _mm_sub_epi8 (sum, res); //actual rounding compensation
 }
 
-uint16x8_t vhaddq_u16(uint16x8_t a, uint16x8_t b); // VHADD.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vhaddq_u16(uint16x8_t a, uint16x8_t b); // VHADD.s16 q0,q0,q0
 _NEON2SSE_INLINE uint16x8_t vhaddq_u16(uint16x8_t a, uint16x8_t b) // VHADD.s16 q0,q0,q0
 {
     __m128i sum, res;
@@ -3082,7 +3084,7 @@ _NEON2SSE_INLINE uint16x8_t vhaddq_u16(uint16x8_t a, uint16x8_t b) // VHADD.s16
     return _mm_sub_epi16 (sum, res); //actual rounding compensation
 }
 
-uint32x4_t vhaddq_u32(uint32x4_t a, uint32x4_t b); // VHADD.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vhaddq_u32(uint32x4_t a, uint32x4_t b); // VHADD.U32 q0,q0,q0
 _NEON2SSE_INLINE uint32x4_t vhaddq_u32(uint32x4_t a, uint32x4_t b) // VHADD.U32 q0,q0,q0
 {
     //need to avoid internal overflow, will use the (x&y)+((x^y)>>1).
@@ -3095,7 +3097,7 @@ _NEON2SSE_INLINE uint32x4_t vhaddq_u32(uint32x4_t a, uint32x4_t b) // VHADD.U32
 
 //************************Vector rounding halving add: vrhadd{q}_<type>. Vr[i]:=(Va[i]+Vb[i]+1)>>1   ***************************
 //*****************************************************************************************************************************
-int8x8_t vrhadd_s8(int8x8_t a,  int8x8_t b); // VRHADD.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vrhadd_s8(int8x8_t a,  int8x8_t b); // VRHADD.S8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vrhadd_s8(int8x8_t a,  int8x8_t b)
 {
     int8x8_t res64;
@@ -3103,7 +3105,7 @@ _NEON2SSE_INLINE int8x8_t vrhadd_s8(int8x8_t a,  int8x8_t b)
 }
 
 
-int16x4_t vrhadd_s16(int16x4_t a,  int16x4_t b); // VRHADD.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vrhadd_s16(int16x4_t a,  int16x4_t b); // VRHADD.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vrhadd_s16(int16x4_t a,  int16x4_t b)
 {
     int16x4_t res64;
@@ -3111,7 +3113,7 @@ _NEON2SSE_INLINE int16x4_t vrhadd_s16(int16x4_t a,  int16x4_t b)
 }
 
 
-int32x2_t vrhadd_s32(int32x2_t a,  int32x2_t b); // VRHADD.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vrhadd_s32(int32x2_t a,  int32x2_t b); // VRHADD.S32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vrhadd_s32(int32x2_t a,  int32x2_t b)
 {
     int32x2_t res64;
@@ -3119,7 +3121,7 @@ _NEON2SSE_INLINE int32x2_t vrhadd_s32(int32x2_t a,  int32x2_t b)
 }
 
 
-uint8x8_t vrhadd_u8(uint8x8_t a, uint8x8_t b); // VRHADD.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vrhadd_u8(uint8x8_t a, uint8x8_t b); // VRHADD.U8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vrhadd_u8(uint8x8_t a, uint8x8_t b)
 {
     uint8x8_t res64;
@@ -3127,7 +3129,7 @@ _NEON2SSE_INLINE uint8x8_t vrhadd_u8(uint8x8_t a, uint8x8_t b)
 }
 
 
-uint16x4_t vrhadd_u16(uint16x4_t a, uint16x4_t b); // VRHADD.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vrhadd_u16(uint16x4_t a, uint16x4_t b); // VRHADD.s16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vrhadd_u16(uint16x4_t a, uint16x4_t b)
 {
     uint16x4_t res64;
@@ -3135,7 +3137,7 @@ _NEON2SSE_INLINE uint16x4_t vrhadd_u16(uint16x4_t a, uint16x4_t b)
 }
 
 
-uint32x2_t vrhadd_u32(uint32x2_t a,  uint32x2_t b); // VRHADD.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vrhadd_u32(uint32x2_t a,  uint32x2_t b); // VRHADD.U32 d0,d0,d0
 _NEON2SSE_INLINE uint32x2_t vrhadd_u32(uint32x2_t a,  uint32x2_t b)
 {
     uint32x2_t res64;
@@ -3143,7 +3145,7 @@ _NEON2SSE_INLINE uint32x2_t vrhadd_u32(uint32x2_t a,  uint32x2_t b)
 }
 
 
-int8x16_t  vrhaddq_s8(int8x16_t a, int8x16_t b); // VRHADD.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t  vrhaddq_s8(int8x16_t a, int8x16_t b); // VRHADD.S8 q0,q0,q0
 _NEON2SSE_INLINE int8x16_t  vrhaddq_s8(int8x16_t a, int8x16_t b) // VRHADD.S8 q0,q0,q0
 {
     //no signed average in x86 SIMD, go to unsigned
@@ -3155,7 +3157,7 @@ _NEON2SSE_INLINE int8x16_t  vrhaddq_s8(int8x16_t a, int8x16_t b) // VRHADD.S8 q0
     return _mm_add_epi8 (sum, c128); //sub 128
 }
 
-int16x8_t  vrhaddq_s16(int16x8_t a, int16x8_t b); // VRHADD.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t  vrhaddq_s16(int16x8_t a, int16x8_t b); // VRHADD.S16 q0,q0,q0
 _NEON2SSE_INLINE int16x8_t  vrhaddq_s16(int16x8_t a, int16x8_t b) // VRHADD.S16 q0,q0,q0
 {
     //no signed average in x86 SIMD, go to unsigned
@@ -3167,7 +3169,7 @@ _NEON2SSE_INLINE int16x8_t  vrhaddq_s16(int16x8_t a, int16x8_t b) // VRHADD.S16
     return _mm_add_epi16 (sum, cx8000); //sub 32768
 }
 
-int32x4_t  vrhaddq_s32(int32x4_t a, int32x4_t b); // VRHADD.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t  vrhaddq_s32(int32x4_t a, int32x4_t b); // VRHADD.S32 q0,q0,q0
 _NEON2SSE_INLINE int32x4_t  vrhaddq_s32(int32x4_t a, int32x4_t b)
 {
     //need to avoid overflow
@@ -3181,14 +3183,14 @@ _NEON2SSE_INLINE int32x4_t  vrhaddq_s32(int32x4_t a, int32x4_t b)
     return _mm_add_epi32(sum,res);
 }
 
-uint8x16_t   vrhaddq_u8(uint8x16_t a, uint8x16_t b); // VRHADD.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t   vrhaddq_u8(uint8x16_t a, uint8x16_t b); // VRHADD.U8 q0,q0,q0
 #define vrhaddq_u8 _mm_avg_epu8 //SSE2, results rounded
 
-uint16x8_t   vrhaddq_u16(uint16x8_t a, uint16x8_t b); // VRHADD.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   vrhaddq_u16(uint16x8_t a, uint16x8_t b); // VRHADD.s16 q0,q0,q0
 #define vrhaddq_u16 _mm_avg_epu16 //SSE2, results rounded
 
 
-uint32x4_t vrhaddq_u32(uint32x4_t a, uint32x4_t b); // VRHADD.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vrhaddq_u32(uint32x4_t a, uint32x4_t b); // VRHADD.U32 q0,q0,q0
 _NEON2SSE_INLINE uint32x4_t vrhaddq_u32(uint32x4_t a, uint32x4_t b) // VRHADD.U32 q0,q0,q0
 {
     //need to avoid overflow
@@ -3204,7 +3206,7 @@ _NEON2SSE_INLINE uint32x4_t vrhaddq_u32(uint32x4_t a, uint32x4_t b) // VRHADD.U3
 
 //****************** VQADD: Vector saturating add ************************
 //************************************************************************
-int8x8_t vqadd_s8(int8x8_t a, int8x8_t b); // VQADD.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vqadd_s8(int8x8_t a, int8x8_t b); // VQADD.S8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vqadd_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
@@ -3212,7 +3214,7 @@ _NEON2SSE_INLINE int8x8_t vqadd_s8(int8x8_t a, int8x8_t b)
 }
 
 
-int16x4_t vqadd_s16(int16x4_t a, int16x4_t b); // VQADD.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vqadd_s16(int16x4_t a, int16x4_t b); // VQADD.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vqadd_s16(int16x4_t a, int16x4_t b)
 {
     int16x4_t res64;
@@ -3220,7 +3222,7 @@ _NEON2SSE_INLINE int16x4_t vqadd_s16(int16x4_t a, int16x4_t b)
 }
 
 
-int32x2_t vqadd_s32(int32x2_t a,  int32x2_t b); // VQADD.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqadd_s32(int32x2_t a,  int32x2_t b); // VQADD.S32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vqadd_s32(int32x2_t a,  int32x2_t b)
 {
     int32x2_t res64;
@@ -3228,7 +3230,7 @@ _NEON2SSE_INLINE int32x2_t vqadd_s32(int32x2_t a,  int32x2_t b)
 }
 
 
-int64x1_t  vqadd_s64(int64x1_t a, int64x1_t b); // VQADD.S64 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t  vqadd_s64(int64x1_t a, int64x1_t b); // VQADD.S64 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vqadd_s64(int64x1_t a, int64x1_t b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int64x1_t res;
@@ -3243,7 +3245,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vqadd_s64(int64x1_t a,
     return res;
 }
 
-uint8x8_t vqadd_u8(uint8x8_t a, uint8x8_t b); // VQADD.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vqadd_u8(uint8x8_t a, uint8x8_t b); // VQADD.U8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vqadd_u8(uint8x8_t a, uint8x8_t b)
 {
     uint8x8_t res64;
@@ -3251,7 +3253,7 @@ _NEON2SSE_INLINE uint8x8_t vqadd_u8(uint8x8_t a, uint8x8_t b)
 }
 
 
-uint16x4_t vqadd_u16(uint16x4_t a, uint16x4_t b); // VQADD.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vqadd_u16(uint16x4_t a, uint16x4_t b); // VQADD.s16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vqadd_u16(uint16x4_t a, uint16x4_t b)
 {
     uint16x4_t res64;
@@ -3259,7 +3261,7 @@ _NEON2SSE_INLINE uint16x4_t vqadd_u16(uint16x4_t a, uint16x4_t b)
 }
 
 
-uint32x2_t vqadd_u32(uint32x2_t a,  uint32x2_t b); // VQADD.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vqadd_u32(uint32x2_t a,  uint32x2_t b); // VQADD.U32 d0,d0,d0
 _NEON2SSE_INLINE uint32x2_t vqadd_u32(uint32x2_t a,  uint32x2_t b)
 {
     uint32x2_t res64;
@@ -3267,7 +3269,7 @@ _NEON2SSE_INLINE uint32x2_t vqadd_u32(uint32x2_t a,  uint32x2_t b)
 }
 
 
-uint64x1_t vqadd_u64(uint64x1_t a, uint64x1_t b); // VQADD.U64 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vqadd_u64(uint64x1_t a, uint64x1_t b); // VQADD.U64 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vqadd_u64(uint64x1_t a, uint64x1_t b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     _NEON2SSE_ALIGN_16 uint64_t a64, b64;
@@ -3281,13 +3283,13 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vqadd_u64(uint64x1_t a
     return res;
 }
 
-int8x16_t   vqaddq_s8(int8x16_t a, int8x16_t b); // VQADD.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t   vqaddq_s8(int8x16_t a, int8x16_t b); // VQADD.S8 q0,q0,q0
 #define vqaddq_s8 _mm_adds_epi8
 
-int16x8_t   vqaddq_s16(int16x8_t a, int16x8_t b); // VQADD.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t   vqaddq_s16(int16x8_t a, int16x8_t b); // VQADD.S16 q0,q0,q0
 #define vqaddq_s16 _mm_adds_epi16
 
-int32x4_t  vqaddq_s32(int32x4_t a, int32x4_t b); // VQADD.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t  vqaddq_s32(int32x4_t a, int32x4_t b); // VQADD.S32 q0,q0,q0
 _NEON2SSE_INLINE int32x4_t  vqaddq_s32(int32x4_t a, int32x4_t b)
 {
     //no corresponding x86 SIMD soulution, special tricks are necessary. Overflow happens only if a and b have the same sign and sum has the opposite sign
@@ -3305,7 +3307,7 @@ _NEON2SSE_INLINE int32x4_t  vqaddq_s32(int32x4_t a, int32x4_t b)
     return _mm_or_si128(res, res_sat);
 }
 
-int64x2_t  vqaddq_s64(int64x2_t a, int64x2_t b); // VQADD.S64 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t  vqaddq_s64(int64x2_t a, int64x2_t b); // VQADD.S64 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqaddq_s64(int64x2_t a, int64x2_t b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     _NEON2SSE_ALIGN_16 uint64_t atmp[2], btmp[2], res[2];
@@ -3326,13 +3328,13 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqaddq_s64(int64x2_t a,
     return _mm_load_si128((__m128i*)res);
 }
 
-uint8x16_t   vqaddq_u8(uint8x16_t a, uint8x16_t b); // VQADD.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t   vqaddq_u8(uint8x16_t a, uint8x16_t b); // VQADD.U8 q0,q0,q0
 #define vqaddq_u8 _mm_adds_epu8
 
-uint16x8_t   vqaddq_u16(uint16x8_t a, uint16x8_t b); // VQADD.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   vqaddq_u16(uint16x8_t a, uint16x8_t b); // VQADD.s16 q0,q0,q0
 #define vqaddq_u16 _mm_adds_epu16
 
-uint32x4_t vqaddq_u32(uint32x4_t a, uint32x4_t b); // VQADD.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vqaddq_u32(uint32x4_t a, uint32x4_t b); // VQADD.U32 q0,q0,q0
 _NEON2SSE_INLINE uint32x4_t vqaddq_u32(uint32x4_t a, uint32x4_t b)
 {
     __m128i c80000000, cmp, subsum, suba, sum;
@@ -3344,7 +3346,7 @@ _NEON2SSE_INLINE uint32x4_t vqaddq_u32(uint32x4_t a, uint32x4_t b)
     return _mm_or_si128 (sum, cmp); //saturation
 }
 
-uint64x2_t vqaddq_u64(uint64x2_t a, uint64x2_t b); // VQADD.U64 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vqaddq_u64(uint64x2_t a, uint64x2_t b); // VQADD.U64 q0,q0,q0
 #ifdef USE_SSE4
     _NEON2SSE_INLINE uint64x2_t vqaddq_u64(uint64x2_t a, uint64x2_t b)
     {
@@ -3373,7 +3375,7 @@ uint64x2_t vqaddq_u64(uint64x2_t a, uint64x2_t b); // VQADD.U64 q0,q0,q0
 
 //******************* Vector add high half (truncated)  ******************
 //************************************************************************
-int8x8_t   vaddhn_s16(int16x8_t a, int16x8_t b); // VADDHN.I16 d0,q0,q0
+_NEON2SSESTORAGE int8x8_t   vaddhn_s16(int16x8_t a, int16x8_t b); // VADDHN.I16 d0,q0,q0
 _NEON2SSE_INLINE int8x8_t   vaddhn_s16(int16x8_t a, int16x8_t b) // VADDHN.I16 d0,q0,q0
 {
     int8x8_t res64;
@@ -3384,7 +3386,7 @@ _NEON2SSE_INLINE int8x8_t   vaddhn_s16(int16x8_t a, int16x8_t b) // VADDHN.I16 d
     return64(sum);
 }
 
-int16x4_t  vaddhn_s32(int32x4_t a, int32x4_t b); // VADDHN.I32 d0,q0,q0
+_NEON2SSESTORAGE int16x4_t  vaddhn_s32(int32x4_t a, int32x4_t b); // VADDHN.I32 d0,q0,q0
 _NEON2SSE_INLINE int16x4_t  vaddhn_s32(int32x4_t a, int32x4_t b) // VADDHN.I32 d0,q0,q0
 {
     int16x4_t res64;
@@ -3395,7 +3397,7 @@ _NEON2SSE_INLINE int16x4_t  vaddhn_s32(int32x4_t a, int32x4_t b) // VADDHN.I32 d
     return64(sum);
 }
 
-int32x2_t  vaddhn_s64(int64x2_t a, int64x2_t b); // VADDHN.I64 d0,q0,q0
+_NEON2SSESTORAGE int32x2_t  vaddhn_s64(int64x2_t a, int64x2_t b); // VADDHN.I64 d0,q0,q0
 _NEON2SSE_INLINE int32x2_t  vaddhn_s64(int64x2_t a, int64x2_t b)
 {
     int32x2_t res64;
@@ -3405,7 +3407,7 @@ _NEON2SSE_INLINE int32x2_t  vaddhn_s64(int64x2_t a, int64x2_t b)
     return64(sum);
 }
 
-uint8x8_t  vaddhn_u16(uint16x8_t a, uint16x8_t b); // VADDHN.I16 d0,q0,q0
+_NEON2SSESTORAGE uint8x8_t  vaddhn_u16(uint16x8_t a, uint16x8_t b); // VADDHN.I16 d0,q0,q0
 _NEON2SSE_INLINE uint8x8_t  vaddhn_u16(uint16x8_t a, uint16x8_t b) // VADDHN.I16 d0,q0,q0
 {
     uint8x8_t res64;
@@ -3416,27 +3418,27 @@ _NEON2SSE_INLINE uint8x8_t  vaddhn_u16(uint16x8_t a, uint16x8_t b) // VADDHN.I16
     return64(sum);
 }
 
-uint16x4_t vaddhn_u32(uint32x4_t a, uint32x4_t b); // VADDHN.I32 d0,q0,q0
+_NEON2SSESTORAGE uint16x4_t vaddhn_u32(uint32x4_t a, uint32x4_t b); // VADDHN.I32 d0,q0,q0
 _NEON2SSE_INLINE uint16x4_t vaddhn_u32(uint32x4_t a, uint32x4_t b) // VADDHN.I32 d0,q0,q0
 {
     uint16x4_t res64;
      __m128i sum;
     sum = _mm_add_epi32 (a, b);
     sum = _mm_srli_epi32 (sum, 16);
-     #ifdef USE_SSE4
-        sum = _MM_PACKUS1_EPI32 (sum); //use 64 low bits only
-    #else
-        sum = _mm_shuffle_epi8 (sum, *(__m128i*) mask8_32_even_odd); //go to 16 bits 
-    #endif
+#ifdef USE_SSE4
+    sum = _MM_PACKUS1_EPI32 (sum); //use 64 low bits only
+#else
+    sum = _mm_shuffle_epi8 (sum, *(__m128i*) mask8_32_even_odd); //go to 16 bits
+#endif
     return64(sum);
 }
 
-uint32x2_t vaddhn_u64(uint64x2_t a, uint64x2_t b); // VADDHN.I64 d0,q0,q0
+_NEON2SSESTORAGE uint32x2_t vaddhn_u64(uint64x2_t a, uint64x2_t b); // VADDHN.I64 d0,q0,q0
 #define vaddhn_u64 vaddhn_s64
 
 //*********** Vector rounding add high half: vraddhn_<type> ******************.
 //***************************************************************************
-int8x8_t   vraddhn_s16(int16x8_t a, int16x8_t b); // VRADDHN.I16 d0,q0,q0
+_NEON2SSESTORAGE int8x8_t   vraddhn_s16(int16x8_t a, int16x8_t b); // VRADDHN.I16 d0,q0,q0
 _NEON2SSE_INLINE int8x8_t   vraddhn_s16(int16x8_t a, int16x8_t b) // VRADDHN.I16 d0,q0,q0
 {
     int8x8_t res64;
@@ -3450,7 +3452,7 @@ _NEON2SSE_INLINE int8x8_t   vraddhn_s16(int16x8_t a, int16x8_t b) // VRADDHN.I16
     return64(sum);
 }
 
-int16x4_t  vraddhn_s32(int32x4_t a, int32x4_t b); // VRADDHN.I32 d0,q0,q0
+_NEON2SSESTORAGE int16x4_t  vraddhn_s32(int32x4_t a, int32x4_t b); // VRADDHN.I32 d0,q0,q0
 _NEON2SSE_INLINE int16x4_t  vraddhn_s32(int32x4_t a, int32x4_t b) // VRADDHN.I32 d0,q0,q0
 {
     //SIMD may be not optimal, serial may be faster
@@ -3465,7 +3467,7 @@ _NEON2SSE_INLINE int16x4_t  vraddhn_s32(int32x4_t a, int32x4_t b) // VRADDHN.I32
     return64(sum);
 }
 
-int32x2_t  vraddhn_s64(int64x2_t a, int64x2_t b); // VRADDHN.I64 d0,q0,q0
+_NEON2SSESTORAGE int32x2_t  vraddhn_s64(int64x2_t a, int64x2_t b); // VRADDHN.I64 d0,q0,q0
 _NEON2SSE_INLINE int32x2_t vraddhn_s64(int64x2_t a, int64x2_t b)
 {
     //SIMD may be not optimal, serial may be faster
@@ -3479,7 +3481,7 @@ _NEON2SSE_INLINE int32x2_t vraddhn_s64(int64x2_t a, int64x2_t b)
     return64(sum);
 }
 
-uint8x8_t  vraddhn_u16(uint16x8_t a, uint16x8_t b); // VRADDHN.I16 d0,q0,q0
+_NEON2SSESTORAGE uint8x8_t  vraddhn_u16(uint16x8_t a, uint16x8_t b); // VRADDHN.I16 d0,q0,q0
 _NEON2SSE_INLINE uint8x8_t  vraddhn_u16(uint16x8_t a, uint16x8_t b) // VRADDHN.I16 d0,q0,q0
 {
     uint8x8_t res64;
@@ -3493,7 +3495,7 @@ _NEON2SSE_INLINE uint8x8_t  vraddhn_u16(uint16x8_t a, uint16x8_t b) // VRADDHN.I
     return64(sum);
 }
 
-uint16x4_t vraddhn_u32(uint32x4_t a, uint32x4_t b); // VRADDHN.I32 d0,q0,q0
+_NEON2SSESTORAGE uint16x4_t vraddhn_u32(uint32x4_t a, uint32x4_t b); // VRADDHN.I32 d0,q0,q0
 _NEON2SSE_INLINE uint16x4_t vraddhn_u32(uint32x4_t a, uint32x4_t b)
 {
     //SIMD may be not optimal, serial may be faster
@@ -3508,7 +3510,7 @@ _NEON2SSE_INLINE uint16x4_t vraddhn_u32(uint32x4_t a, uint32x4_t b)
     return64(sum);
 }
 
-uint32x2_t vraddhn_u64(uint64x2_t a, uint64x2_t b); // VRADDHN.I64 d0,q0,q0
+_NEON2SSESTORAGE uint32x2_t vraddhn_u64(uint64x2_t a, uint64x2_t b); // VRADDHN.I64 d0,q0,q0
 #define vraddhn_u64 vraddhn_s64
 
 //**********************************************************************************
@@ -3517,7 +3519,7 @@ uint32x2_t vraddhn_u64(uint64x2_t a, uint64x2_t b); // VRADDHN.I64 d0,q0,q0
 
 //Vector multiply: vmul -> Vr[i] := Va[i] * Vb[i]
 //As we don't go to wider result functions are equal to "multiply low" in x86
-int8x8_t vmul_s8(int8x8_t a, int8x8_t b); // VMUL.I8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vmul_s8(int8x8_t a, int8x8_t b); // VMUL.I8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vmul_s8(int8x8_t a, int8x8_t b) // VMUL.I8 d0,d0,d0
 {
     // no 8 bit simd multiply, need to go to 16 bits in SSE
@@ -3530,13 +3532,13 @@ _NEON2SSE_INLINE int8x8_t vmul_s8(int8x8_t a, int8x8_t b) // VMUL.I8 d0,d0,d0
     return64(res);
 }
 
-int16x4_t vmul_s16(int16x4_t a,  int16x4_t b); // VMUL.I16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vmul_s16(int16x4_t a,  int16x4_t b); // VMUL.I16 d0,d0,d0
 #define vmul_s16 vmul_u16
 
-int32x2_t vmul_s32(int32x2_t a,  int32x2_t b); // VMUL.I32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vmul_s32(int32x2_t a,  int32x2_t b); // VMUL.I32 d0,d0,d0
 #define vmul_s32 vmul_u32
 
-float32x2_t vmul_f32(float32x2_t a, float32x2_t b); // VMUL.F32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vmul_f32(float32x2_t a, float32x2_t b); // VMUL.F32 d0,d0,d0
 _NEON2SSE_INLINE float32x2_t vmul_f32(float32x2_t a, float32x2_t b)
 {
     float32x4_t tmp;
@@ -3546,7 +3548,7 @@ _NEON2SSE_INLINE float32x2_t vmul_f32(float32x2_t a, float32x2_t b)
     return res64;
 }
 
-uint8x8_t vmul_u8(uint8x8_t a, uint8x8_t b); // VMUL.I8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vmul_u8(uint8x8_t a, uint8x8_t b); // VMUL.I8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vmul_u8(uint8x8_t a, uint8x8_t b) // VMUL.I8 d0,d0,d0
 {
     // no 8 bit simd multiply, need to go to 16 bits in SSE
@@ -3561,14 +3563,14 @@ _NEON2SSE_INLINE uint8x8_t vmul_u8(uint8x8_t a, uint8x8_t b) // VMUL.I8 d0,d0,d0
     return64(res);
 }
 
-uint16x4_t vmul_u16(uint16x4_t a, uint16x4_t b); // VMUL.I16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vmul_u16(uint16x4_t a, uint16x4_t b); // VMUL.I16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vmul_u16(uint16x4_t a, uint16x4_t b)
 {
     uint16x4_t res64;
     return64(_mm_mullo_epi16(_pM128i(a),_pM128i(b)));
 }
 
-uint32x2_t   vmul_u32(uint32x2_t a, uint32x2_t b); // VMUL.I32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t   vmul_u32(uint32x2_t a, uint32x2_t b); // VMUL.I32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING( uint32x2_t   vmul_u32(uint32x2_t a, uint32x2_t b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     uint32x2_t res;
@@ -3577,7 +3579,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING( uint32x2_t   vmul_u32(uint32x2_t
     return res;
 }
 
-poly8x8_t vmul_p8(poly8x8_t a, poly8x8_t b); // VMUL.P8 d0,d0,d0
+_NEON2SSESTORAGE poly8x8_t vmul_p8(poly8x8_t a, poly8x8_t b); // VMUL.P8 d0,d0,d0
 _NEON2SSE_INLINE poly8x8_t vmul_p8(poly8x8_t a, poly8x8_t b)
 {
     //may be optimized
@@ -3599,7 +3601,7 @@ _NEON2SSE_INLINE poly8x8_t vmul_p8(poly8x8_t a, poly8x8_t b)
     return64 (res);
 }
 
-int8x16_t vmulq_s8(int8x16_t a, int8x16_t b); // VMUL.I8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vmulq_s8(int8x16_t a, int8x16_t b); // VMUL.I8 q0,q0,q0
 _NEON2SSE_INLINE int8x16_t vmulq_s8(int8x16_t a, int8x16_t b) // VMUL.I8 q0,q0,q0
 {
     // no 8 bit simd multiply, need to go to 16 bits
@@ -3621,16 +3623,16 @@ _NEON2SSE_INLINE int8x16_t vmulq_s8(int8x16_t a, int8x16_t b) // VMUL.I8 q0,q0,q
     return _mm_unpacklo_epi64(r16_1,  r16_2);
 }
 
-int16x8_t   vmulq_s16(int16x8_t a, int16x8_t b); // VMUL.I16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t   vmulq_s16(int16x8_t a, int16x8_t b); // VMUL.I16 q0,q0,q0
 #define vmulq_s16 _mm_mullo_epi16
 
-int32x4_t   vmulq_s32(int32x4_t a, int32x4_t b); // VMUL.I32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t   vmulq_s32(int32x4_t a, int32x4_t b); // VMUL.I32 q0,q0,q0
 #define vmulq_s32 _MM_MULLO_EPI32 //SSE4.1
 
-float32x4_t vmulq_f32(float32x4_t a, float32x4_t b); // VMUL.F32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vmulq_f32(float32x4_t a, float32x4_t b); // VMUL.F32 q0,q0,q0
 #define vmulq_f32 _mm_mul_ps
 
-uint8x16_t vmulq_u8(uint8x16_t a, uint8x16_t b); // VMUL.I8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vmulq_u8(uint8x16_t a, uint8x16_t b); // VMUL.I8 q0,q0,q0
 _NEON2SSE_INLINE uint8x16_t vmulq_u8(uint8x16_t a, uint8x16_t b) // VMUL.I8 q0,q0,q0
 {
     // no 8 bit simd multiply, need to go to 16 bits
@@ -3652,13 +3654,13 @@ _NEON2SSE_INLINE uint8x16_t vmulq_u8(uint8x16_t a, uint8x16_t b) // VMUL.I8 q0,q
     return _mm_packus_epi16 (r16_1,  r16_2);
 }
 
-uint16x8_t   vmulq_u16(uint16x8_t a, uint16x8_t b); // VMUL.I16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   vmulq_u16(uint16x8_t a, uint16x8_t b); // VMUL.I16 q0,q0,q0
 #define vmulq_u16 _mm_mullo_epi16
 
-uint32x4_t   vmulq_u32(uint32x4_t a, uint32x4_t b); // VMUL.I32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t   vmulq_u32(uint32x4_t a, uint32x4_t b); // VMUL.I32 q0,q0,q0
 #define vmulq_u32 _MM_MULLO_EPI32 //SSE4.1
 
-poly8x16_t vmulq_p8(poly8x16_t a, poly8x16_t b); // VMUL.P8 q0,q0,q0
+_NEON2SSESTORAGE poly8x16_t vmulq_p8(poly8x16_t a, poly8x16_t b); // VMUL.P8 q0,q0,q0
 _NEON2SSE_INLINE poly8x16_t vmulq_p8(poly8x16_t a, poly8x16_t b)
 {
     //may be optimized
@@ -3679,7 +3681,7 @@ _NEON2SSE_INLINE poly8x16_t vmulq_p8(poly8x16_t a, poly8x16_t b)
 
 //************************* Vector long multiply ***********************************
 //****************************************************************************
-int16x8_t vmull_s8(int8x8_t a, int8x8_t b); // VMULL.S8 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vmull_s8(int8x8_t a, int8x8_t b); // VMULL.S8 q0,d0,d0
 _NEON2SSE_INLINE int16x8_t vmull_s8(int8x8_t a, int8x8_t b) // VMULL.S8 q0,d0,d0
 {
     //no 8 bit simd multiply, need to go to 16 bits
@@ -3689,25 +3691,25 @@ _NEON2SSE_INLINE int16x8_t vmull_s8(int8x8_t a, int8x8_t b) // VMULL.S8 q0,d0,d0
     return _mm_mullo_epi16 (a16, b16); //should fit into 16 bit
 }
 
-int32x4_t vmull_s16(int16x4_t a, int16x4_t b); // VMULL.S16 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vmull_s16(int16x4_t a, int16x4_t b); // VMULL.S16 q0,d0,d0
 _NEON2SSE_INLINE int32x4_t vmull_s16(int16x4_t a, int16x4_t b) // VMULL.S16 q0,d0,d0
 {
-    #ifdef USE_SSE4
-        __m128i a16, b16;
-        a16 = _MM_CVTEPI16_EPI32 (_pM128i(a)); // SSE 4.1
-        b16 = _MM_CVTEPI16_EPI32 (_pM128i(b)); // SSE 4.1
-        return _MM_MULLO_EPI32 (a16, b16); // SSE 4.1
-    #else
-        __m128i low, hi, a128,b128;
-        a128 = _pM128i(a);
-        b128 = _pM128i(b);
-        low =  _mm_mullo_epi16(a128,b128);
-        hi =   _mm_mulhi_epi16(a128,b128);
-        return _mm_unpacklo_epi16(low,hi);
-    #endif
+#ifdef USE_SSE4
+    __m128i a16, b16;
+    a16 = _MM_CVTEPI16_EPI32 (_pM128i(a)); // SSE 4.1
+    b16 = _MM_CVTEPI16_EPI32 (_pM128i(b)); // SSE 4.1
+    return _MM_MULLO_EPI32 (a16, b16); // SSE 4.1
+#else
+    __m128i low, hi, a128,b128;
+    a128 = _pM128i(a);
+    b128 = _pM128i(b);
+    low =  _mm_mullo_epi16(a128,b128);
+    hi =   _mm_mulhi_epi16(a128,b128);
+    return _mm_unpacklo_epi16(low,hi);
+#endif
 }
 
-int64x2_t vmull_s32(int32x2_t a, int32x2_t b); // VMULL.S32 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vmull_s32(int32x2_t a, int32x2_t b); // VMULL.S32 q0,d0,d0
 _NEON2SSE_INLINE int64x2_t vmull_s32(int32x2_t a, int32x2_t b) // VMULL.S32 q0,d0,d0
 {
     __m128i ab, ba, a128, b128;
@@ -3718,7 +3720,7 @@ _NEON2SSE_INLINE int64x2_t vmull_s32(int32x2_t a, int32x2_t b) // VMULL.S32 q0,d
     return _MM_MUL_EPI32(ab, ba); //uses 1rst and 3rd data lanes, the multiplication gives 64 bit result
 }
 
-uint16x8_t vmull_u8(uint8x8_t a, uint8x8_t b); // VMULL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vmull_u8(uint8x8_t a, uint8x8_t b); // VMULL.U8 q0,d0,d0
 _NEON2SSE_INLINE uint16x8_t vmull_u8(uint8x8_t a, uint8x8_t b) // VMULL.U8 q0,d0,d0
 {
     //no 8 bit simd multiply, need to go to 16 bits
@@ -3728,25 +3730,25 @@ _NEON2SSE_INLINE uint16x8_t vmull_u8(uint8x8_t a, uint8x8_t b) // VMULL.U8 q0,d0
     return _mm_mullo_epi16 (a16, b16); //should fit into 16 bit
 }
 
-uint32x4_t vmull_u16(uint16x4_t a, uint16x4_t b); // VMULL.s16 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vmull_u16(uint16x4_t a, uint16x4_t b); // VMULL.s16 q0,d0,d0
 _NEON2SSE_INLINE uint32x4_t vmull_u16(uint16x4_t a, uint16x4_t b) // VMULL.s16 q0,d0,d0
 {
-    #ifdef USE_SSE4
-        __m128i a16, b16;
-        a16 = _MM_CVTEPU16_EPI32 (_pM128i(a)); // SSE 4.1
-        b16 = _MM_CVTEPU16_EPI32 (_pM128i(b)); // SSE 4.1
-        return _MM_MULLO_EPI32 (a16, b16); // SSE 4.1
-    #else
-        __m128i a128,b128,low, hi;
-        a128 = _pM128i(a);
-        b128 = _pM128i(b);
-        low =  _mm_mullo_epi16(a128,b128);
-        hi =   _mm_mulhi_epu16(a128,b128);
-        return _mm_unpacklo_epi16(low,hi);
-    #endif
+#ifdef USE_SSE4
+    __m128i a16, b16;
+    a16 = _MM_CVTEPU16_EPI32 (_pM128i(a)); // SSE 4.1
+    b16 = _MM_CVTEPU16_EPI32 (_pM128i(b)); // SSE 4.1
+    return _MM_MULLO_EPI32 (a16, b16); // SSE 4.1
+#else
+    __m128i a128,b128,low, hi;
+    a128 = _pM128i(a);
+    b128 = _pM128i(b);
+    low =  _mm_mullo_epi16(a128,b128);
+    hi =   _mm_mulhi_epu16(a128,b128);
+    return _mm_unpacklo_epi16(low,hi);
+#endif
 }
 
-uint64x2_t vmull_u32(uint32x2_t a, uint32x2_t b); // VMULL.U32 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vmull_u32(uint32x2_t a, uint32x2_t b); // VMULL.U32 q0,d0,d0
 _NEON2SSE_INLINE uint64x2_t vmull_u32(uint32x2_t a, uint32x2_t b) // VMULL.U32 q0,d0,d0
 {
     ///may be not optimal compared with serial implementation
@@ -3758,7 +3760,7 @@ _NEON2SSE_INLINE uint64x2_t vmull_u32(uint32x2_t a, uint32x2_t b) // VMULL.U32 q
     return _mm_mul_epu32 (ab, ba); //uses 1rst and 3rd data lanes, the multiplication gives 64 bit result
 }
 
-poly16x8_t vmull_p8(poly8x8_t a, poly8x8_t b); // VMULL.P8 q0,d0,d0
+_NEON2SSESTORAGE poly16x8_t vmull_p8(poly8x8_t a, poly8x8_t b); // VMULL.P8 q0,d0,d0
 _NEON2SSE_INLINE poly16x8_t vmull_p8(poly8x8_t a, poly8x8_t b)
 {
     //may be optimized
@@ -3785,7 +3787,7 @@ _NEON2SSE_INLINE poly16x8_t vmull_p8(poly8x8_t a, poly8x8_t b)
 
 //****************Vector saturating doubling long multiply **************************
 //*****************************************************************
-int32x4_t vqdmull_s16(int16x4_t a, int16x4_t b); // VQDMULL.S16 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vqdmull_s16(int16x4_t a, int16x4_t b); // VQDMULL.S16 q0,d0,d0
 _NEON2SSE_INLINE int32x4_t vqdmull_s16(int16x4_t a, int16x4_t b)
 {
     //the serial soulution may be faster due to saturation
@@ -3794,7 +3796,7 @@ _NEON2SSE_INLINE int32x4_t vqdmull_s16(int16x4_t a, int16x4_t b)
     return vqd_s32(res);
 }
 
-int64x2_t vqdmull_s32(int32x2_t a, int32x2_t b); // VQDMULL.S32 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vqdmull_s32(int32x2_t a, int32x2_t b); // VQDMULL.S32 q0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmull_s32(int32x2_t a, int32x2_t b),_NEON2SSE_REASON_SLOW_SERIAL)
 {
     //the serial soulution may be faster due to saturation
@@ -3805,7 +3807,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmull_s32(int32x2_t a
 
 //********************* Vector multiply accumulate: vmla -> Vr[i] := Va[i] + Vb[i] * Vc[i]  ************************
 //******************************************************************************************
-int8x8_t vmla_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VMLA.I8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vmla_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VMLA.I8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vmla_s8(int8x8_t a, int8x8_t b, int8x8_t c) // VMLA.I8 d0,d0,d0
 {
     // no 8 bit x86 simd multiply, need to go to 16 bits,  and use the low 64 bits
@@ -3819,7 +3821,7 @@ _NEON2SSE_INLINE int8x8_t vmla_s8(int8x8_t a, int8x8_t b, int8x8_t c) // VMLA.I8
     return64(res);
 }
 
-int16x4_t vmla_s16(int16x4_t a,  int16x4_t b, int16x4_t c); // VMLA.I16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vmla_s16(int16x4_t a,  int16x4_t b, int16x4_t c); // VMLA.I16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vmla_s16(int16x4_t a,  int16x4_t b, int16x4_t c)
 {
     int16x4_t res64;
@@ -3827,7 +3829,7 @@ _NEON2SSE_INLINE int16x4_t vmla_s16(int16x4_t a,  int16x4_t b, int16x4_t c)
 }
 
 
-int32x2_t vmla_s32(int32x2_t a, int32x2_t b, int32x2_t c); // VMLA.I32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vmla_s32(int32x2_t a, int32x2_t b, int32x2_t c); // VMLA.I32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vmla_s32(int32x2_t a, int32x2_t b, int32x2_t c) // VMLA.I32 d0,d0,d0
 {
     int32x2_t res64;
@@ -3837,7 +3839,7 @@ _NEON2SSE_INLINE int32x2_t vmla_s32(int32x2_t a, int32x2_t b, int32x2_t c) // VM
     return64(res);
 }
 
-float32x2_t vmla_f32(float32x2_t a, float32x2_t b, float32x2_t c); // VMLA.F32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vmla_f32(float32x2_t a, float32x2_t b, float32x2_t c); // VMLA.F32 d0,d0,d0
 _NEON2SSE_INLINE float32x2_t vmla_f32(float32x2_t a, float32x2_t b, float32x2_t c)
 {
     //fma is coming soon, but right now:
@@ -3849,7 +3851,7 @@ _NEON2SSE_INLINE float32x2_t vmla_f32(float32x2_t a, float32x2_t b, float32x2_t
     return res64;
 }
 
-uint8x8_t vmla_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VMLA.I8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vmla_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VMLA.I8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vmla_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c) // VMLA.I8 d0,d0,d0
 {
     // no 8 bit x86 simd multiply, need to go to 16 bits,  and use the low 64 bits
@@ -3865,13 +3867,13 @@ _NEON2SSE_INLINE uint8x8_t vmla_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c) // VML
     return64(res);
 }
 
-uint16x4_t vmla_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t c); // VMLA.I16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vmla_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t c); // VMLA.I16 d0,d0,d0
 #define vmla_u16 vmla_s16
 
-uint32x2_t vmla_u32(uint32x2_t a,  uint32x2_t b, uint32x2_t c); // VMLA.I32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vmla_u32(uint32x2_t a,  uint32x2_t b, uint32x2_t c); // VMLA.I32 d0,d0,d0
 #define vmla_u32 vmla_s32
 
-int8x16_t vmlaq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VMLA.I8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vmlaq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VMLA.I8 q0,q0,q0
 _NEON2SSE_INLINE int8x16_t vmlaq_s8(int8x16_t a, int8x16_t b, int8x16_t c) // VMLA.I8 q0,q0,q0
 {
     //solution may be not optimal
@@ -3895,7 +3897,7 @@ _NEON2SSE_INLINE int8x16_t vmlaq_s8(int8x16_t a, int8x16_t b, int8x16_t c) // VM
     return _mm_unpacklo_epi64(r16_1,r16_2);
 }
 
-int16x8_t vmlaq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VMLA.I16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vmlaq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VMLA.I16 q0,q0,q0
 _NEON2SSE_INLINE int16x8_t vmlaq_s16(int16x8_t a, int16x8_t b, int16x8_t c) // VMLA.I16 q0,q0,q0
 {
     __m128i res;
@@ -3903,7 +3905,7 @@ _NEON2SSE_INLINE int16x8_t vmlaq_s16(int16x8_t a, int16x8_t b, int16x8_t c) // V
     return _mm_add_epi16 (res, a);
 }
 
-int32x4_t vmlaq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VMLA.I32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vmlaq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VMLA.I32 q0,q0,q0
 _NEON2SSE_INLINE int32x4_t vmlaq_s32(int32x4_t a, int32x4_t b, int32x4_t c) // VMLA.I32 q0,q0,q0
 {
     __m128i res;
@@ -3911,7 +3913,7 @@ _NEON2SSE_INLINE int32x4_t vmlaq_s32(int32x4_t a, int32x4_t b, int32x4_t c) // V
     return _mm_add_epi32 (res, a);
 }
 
-float32x4_t vmlaq_f32(float32x4_t a, float32x4_t b, float32x4_t c); // VMLA.F32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vmlaq_f32(float32x4_t a, float32x4_t b, float32x4_t c); // VMLA.F32 q0,q0,q0
 _NEON2SSE_INLINE float32x4_t vmlaq_f32(float32x4_t a, float32x4_t b, float32x4_t c) // VMLA.F32 q0,q0,q0
 {
     //fma is coming soon, but right now:
@@ -3920,7 +3922,7 @@ _NEON2SSE_INLINE float32x4_t vmlaq_f32(float32x4_t a, float32x4_t b, float32x4_t
     return _mm_add_ps (a, res);
 }
 
-uint8x16_t vmlaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VMLA.I8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vmlaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VMLA.I8 q0,q0,q0
 _NEON2SSE_INLINE uint8x16_t vmlaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c) // VMLA.I8 q0,q0,q0
 {
     //solution may be not optimal
@@ -3944,16 +3946,16 @@ _NEON2SSE_INLINE uint8x16_t vmlaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c) /
     return _mm_unpacklo_epi64(r16_1,r16_2);
 }
 
-uint16x8_t vmlaq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VMLA.I16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vmlaq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VMLA.I16 q0,q0,q0
 #define vmlaq_u16 vmlaq_s16
 
-uint32x4_t vmlaq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VMLA.I32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vmlaq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VMLA.I32 q0,q0,q0
 #define vmlaq_u32 vmlaq_s32
 
 //**********************  Vector widening multiply accumulate (long multiply accumulate):
 //                          vmla -> Vr[i] := Va[i] + Vb[i] * Vc[i]  **************
 //********************************************************************************************
-int16x8_t vmlal_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VMLAL.S8 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vmlal_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VMLAL.S8 q0,d0,d0
 _NEON2SSE_INLINE int16x8_t vmlal_s8(int16x8_t a, int8x8_t b, int8x8_t c) // VMLAL.S8 q0,d0,d0
 {
     int16x8_t res;
@@ -3961,7 +3963,7 @@ _NEON2SSE_INLINE int16x8_t vmlal_s8(int16x8_t a, int8x8_t b, int8x8_t c) // VMLA
     return _mm_add_epi16 (res, a);
 }
 
-int32x4_t vmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VMLAL.S16 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VMLAL.S16 q0,d0,d0
 _NEON2SSE_INLINE int32x4_t vmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c) // VMLAL.S16 q0,d0,d0
 {
     //may be not optimal compared with serial implementation
@@ -3970,7 +3972,7 @@ _NEON2SSE_INLINE int32x4_t vmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c) // V
     return _mm_add_epi32 (res, a);
 }
 
-int64x2_t vmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VMLAL.S32 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VMLAL.S32 q0,d0,d0
 _NEON2SSE_INLINE int64x2_t vmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c) // VMLAL.S32 q0,d0,d0
 {
     //may be not optimal compared with serial implementation
@@ -3979,7 +3981,7 @@ _NEON2SSE_INLINE int64x2_t vmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c) // V
     return _mm_add_epi64 (res, a);
 }
 
-uint16x8_t vmlal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VMLAL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vmlal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VMLAL.U8 q0,d0,d0
 _NEON2SSE_INLINE uint16x8_t vmlal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c) // VMLAL.U8 q0,d0,d0
 {
     uint16x8_t res;
@@ -3987,7 +3989,7 @@ _NEON2SSE_INLINE uint16x8_t vmlal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c) //
     return _mm_add_epi16 (res, a);
 }
 
-uint32x4_t vmlal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VMLAL.s16 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vmlal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VMLAL.s16 q0,d0,d0
 _NEON2SSE_INLINE uint32x4_t vmlal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c) // VMLAL.s16 q0,d0,d0
 {
     //may be not optimal compared with serial implementation
@@ -3996,7 +3998,7 @@ _NEON2SSE_INLINE uint32x4_t vmlal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c)
     return _mm_add_epi32 (res, a);
 }
 
-uint64x2_t vmlal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VMLAL.U32 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vmlal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VMLAL.U32 q0,d0,d0
 _NEON2SSE_INLINE uint64x2_t vmlal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c) // VMLAL.U32 q0,d0,d0
 {
     //may be not optimal compared with serial implementation
@@ -4007,7 +4009,7 @@ _NEON2SSE_INLINE uint64x2_t vmlal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c)
 
 //******************** Vector multiply subtract: vmls -> Vr[i] := Va[i] - Vb[i] * Vc[i] ***************************************
 //********************************************************************************************
-int8x8_t vmls_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VMLS.I8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vmls_s8(int8x8_t a, int8x8_t b, int8x8_t c); // VMLS.I8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vmls_s8(int8x8_t a, int8x8_t b, int8x8_t c) // VMLS.I8 d0,d0,d0
 {
     // no 8 bit simd multiply, need to go to 16 bits -  and use the low 64 bits
@@ -4018,7 +4020,7 @@ _NEON2SSE_INLINE int8x8_t vmls_s8(int8x8_t a, int8x8_t b, int8x8_t c) // VMLS.I8
     return64(res);
 }
 
-int16x4_t vmls_s16(int16x4_t a,  int16x4_t b, int16x4_t c); // VMLS.I16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vmls_s16(int16x4_t a,  int16x4_t b, int16x4_t c); // VMLS.I16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vmls_s16(int16x4_t a,  int16x4_t b, int16x4_t c)
 {
     int16x4_t res64;
@@ -4026,7 +4028,7 @@ _NEON2SSE_INLINE int16x4_t vmls_s16(int16x4_t a,  int16x4_t b, int16x4_t c)
 }
 
 
-int32x2_t vmls_s32(int32x2_t a, int32x2_t b, int32x2_t c); // VMLS.I32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vmls_s32(int32x2_t a, int32x2_t b, int32x2_t c); // VMLS.I32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vmls_s32(int32x2_t a, int32x2_t b, int32x2_t c) // VMLS.I32 d0,d0,d0
 {
     int32x2_t res64;
@@ -4036,7 +4038,7 @@ _NEON2SSE_INLINE int32x2_t vmls_s32(int32x2_t a, int32x2_t b, int32x2_t c) // VM
     return64(res);
 }
 
-float32x2_t vmls_f32(float32x2_t a, float32x2_t b, float32x2_t c); // VMLS.F32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vmls_f32(float32x2_t a, float32x2_t b, float32x2_t c); // VMLS.F32 d0,d0,d0
 _NEON2SSE_INLINE float32x2_t vmls_f32(float32x2_t a, float32x2_t b, float32x2_t c)
 {
     __m128 res;
@@ -4047,7 +4049,7 @@ _NEON2SSE_INLINE float32x2_t vmls_f32(float32x2_t a, float32x2_t b, float32x2_t
     return res64;
 }
 
-uint8x8_t vmls_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VMLS.I8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vmls_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VMLS.I8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vmls_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c)
 {
     // no 8 bit simd multiply, need to go to 16 bits -  and use the low 64 bits
@@ -4058,14 +4060,14 @@ _NEON2SSE_INLINE uint8x8_t vmls_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c)
     return64(res);
 }
 
-uint16x4_t vmls_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t c); // VMLS.I16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vmls_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t c); // VMLS.I16 d0,d0,d0
 #define vmls_u16 vmls_s16
 
-uint32x2_t vmls_u32(uint32x2_t a,  uint32x2_t b, uint32x2_t c); // VMLS.I32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vmls_u32(uint32x2_t a,  uint32x2_t b, uint32x2_t c); // VMLS.I32 d0,d0,d0
 #define vmls_u32 vmls_s32
 
 
-int8x16_t vmlsq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VMLS.I8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vmlsq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VMLS.I8 q0,q0,q0
 _NEON2SSE_INLINE int8x16_t vmlsq_s8(int8x16_t a, int8x16_t b, int8x16_t c) // VMLS.I8 q0,q0,q0
 {
     //solution may be not optimal
@@ -4089,7 +4091,7 @@ _NEON2SSE_INLINE int8x16_t vmlsq_s8(int8x16_t a, int8x16_t b, int8x16_t c) // VM
     return _mm_unpacklo_epi64(r16_1,r16_2);
 }
 
-int16x8_t vmlsq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VMLS.I16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vmlsq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VMLS.I16 q0,q0,q0
 _NEON2SSE_INLINE int16x8_t vmlsq_s16(int16x8_t a, int16x8_t b, int16x8_t c) // VMLS.I16 q0,q0,q0
 {
     __m128i res;
@@ -4097,7 +4099,7 @@ _NEON2SSE_INLINE int16x8_t vmlsq_s16(int16x8_t a, int16x8_t b, int16x8_t c) // V
     return _mm_sub_epi16 (a, res);
 }
 
-int32x4_t vmlsq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VMLS.I32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vmlsq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VMLS.I32 q0,q0,q0
 _NEON2SSE_INLINE int32x4_t vmlsq_s32(int32x4_t a, int32x4_t b, int32x4_t c) // VMLS.I32 q0,q0,q0
 {
     __m128i res;
@@ -4105,7 +4107,7 @@ _NEON2SSE_INLINE int32x4_t vmlsq_s32(int32x4_t a, int32x4_t b, int32x4_t c) // V
     return _mm_sub_epi32 (a, res);
 }
 
-float32x4_t vmlsq_f32(float32x4_t a, float32x4_t b, float32x4_t c); // VMLS.F32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vmlsq_f32(float32x4_t a, float32x4_t b, float32x4_t c); // VMLS.F32 q0,q0,q0
 _NEON2SSE_INLINE float32x4_t vmlsq_f32(float32x4_t a, float32x4_t b, float32x4_t c) // VMLS.F32 q0,q0,q0
 {
     __m128 res;
@@ -4113,7 +4115,7 @@ _NEON2SSE_INLINE float32x4_t vmlsq_f32(float32x4_t a, float32x4_t b, float32x4_t
     return _mm_sub_ps (a, res);
 }
 
-uint8x16_t vmlsq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VMLS.I8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vmlsq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VMLS.I8 q0,q0,q0
 _NEON2SSE_INLINE uint8x16_t vmlsq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c) // VMLS.I8 q0,q0,q0
 {
     //solution may be not optimal
@@ -4137,15 +4139,15 @@ _NEON2SSE_INLINE uint8x16_t vmlsq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c) /
     return _mm_unpacklo_epi64(r16_1,r16_2);
 }
 
-uint16x8_t vmlsq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VMLS.I16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vmlsq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VMLS.I16 q0,q0,q0
 #define vmlsq_u16 vmlsq_s16
 
-uint32x4_t vmlsq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VMLS.I32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vmlsq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VMLS.I32 q0,q0,q0
 #define vmlsq_u32 vmlsq_s32
 
 //******************** Vector multiply subtract long (widening multiply subtract) ************************************
 //*************************************************************************************************************
-int16x8_t vmlsl_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VMLSL.S8 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vmlsl_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VMLSL.S8 q0,d0,d0
 _NEON2SSE_INLINE int16x8_t vmlsl_s8(int16x8_t a, int8x8_t b, int8x8_t c) // VMLSL.S8 q0,d0,d0
 {
     int16x8_t res;
@@ -4153,7 +4155,7 @@ _NEON2SSE_INLINE int16x8_t vmlsl_s8(int16x8_t a, int8x8_t b, int8x8_t c) // VMLS
     return _mm_sub_epi16 (a, res);
 }
 
-int32x4_t vmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VMLSL.S16 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VMLSL.S16 q0,d0,d0
 _NEON2SSE_INLINE int32x4_t vmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c) // VMLSL.S16 q0,d0,d0
 {
     //may be not optimal compared with serial implementation
@@ -4162,7 +4164,7 @@ _NEON2SSE_INLINE int32x4_t vmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c) // V
     return _mm_sub_epi32 (a, res);
 }
 
-int64x2_t vmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VMLSL.S32 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VMLSL.S32 q0,d0,d0
 _NEON2SSE_INLINE int64x2_t vmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c) // VMLSL.S32 q0,d0,d0
 {
     //may be not optimal compared with serial implementation
@@ -4171,7 +4173,7 @@ _NEON2SSE_INLINE int64x2_t vmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c) // V
     return _mm_sub_epi64 (a, res);
 }
 
-uint16x8_t vmlsl_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VMLSL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vmlsl_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VMLSL.U8 q0,d0,d0
 _NEON2SSE_INLINE uint16x8_t vmlsl_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c) // VMLSL.U8 q0,d0,d0
 {
     uint16x8_t res;
@@ -4179,7 +4181,7 @@ _NEON2SSE_INLINE uint16x8_t vmlsl_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c) //
     return _mm_sub_epi16 (a, res);
 }
 
-uint32x4_t vmlsl_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VMLSL.s16 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vmlsl_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VMLSL.s16 q0,d0,d0
 _NEON2SSE_INLINE uint32x4_t vmlsl_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c) // VMLSL.s16 q0,d0,d0
 {
     //may be not optimal compared with serial implementation
@@ -4188,7 +4190,7 @@ _NEON2SSE_INLINE uint32x4_t vmlsl_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c)
     return _mm_sub_epi32 (a, res);
 }
 
-uint64x2_t vmlsl_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VMLSL.U32 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vmlsl_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VMLSL.U32 q0,d0,d0
 _NEON2SSE_INLINE uint64x2_t vmlsl_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c) // VMLSL.U32 q0,d0,d0
 {
     //may be not optimal compared with serial implementation
@@ -4199,7 +4201,7 @@ _NEON2SSE_INLINE uint64x2_t vmlsl_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c)
 
 //******  Vector saturating doubling multiply high **********************
 //*************************************************************************
-int16x4_t vqdmulh_s16(int16x4_t a,  int16x4_t b); // VQDMULH.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vqdmulh_s16(int16x4_t a,  int16x4_t b); // VQDMULH.S16 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x4_t vqdmulh_s16(int16x4_t a,  int16x4_t b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int16x4_t res;
@@ -4213,13 +4215,13 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x4_t vqdmulh_s16(int16x4_t a
     return res;
 }
 
-int32x2_t vqdmulh_s32(int32x2_t a, int32x2_t b); // VQDMULH.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqdmulh_s32(int32x2_t a, int32x2_t b); // VQDMULH.S32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vqdmulh_s32(int32x2_t a, int32x2_t b) // no multiply high 32 bit SIMD in IA32, so need to do some tricks, serial solution may be faster
 {
     //may be not optimal compared with a serial solution
     int32x2_t res64;
     __m128i mask;
-    _NEON2SSE_ALIGN_16 uint32_t cmask32[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
+    _NEON2SSE_ALIGN_16 static const uint32_t cmask32[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
     int64x2_t mul;
     mul = vmull_s32(a,b);
     mul = _mm_slli_epi64(mul,1); //double the result
@@ -4230,11 +4232,11 @@ _NEON2SSE_INLINE int32x2_t vqdmulh_s32(int32x2_t a, int32x2_t b) // no multiply
     return64(mul);
 }
 
-int16x8_t vqdmulhq_s16(int16x8_t a, int16x8_t b); // VQDMULH.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vqdmulhq_s16(int16x8_t a, int16x8_t b); // VQDMULH.S16 q0,q0,q0
 _NEON2SSE_INLINE int16x8_t vqdmulhq_s16(int16x8_t a, int16x8_t b) // VQDMULH.S16 q0,q0,q0
 {
     __m128i res, res_lo, mask;
-    _NEON2SSE_ALIGN_16 uint16_t cmask[] = {0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000};
+    _NEON2SSE_ALIGN_16 static const uint16_t cmask[] = {0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000};
     res = _mm_mulhi_epi16 (a, b);
     res = _mm_slli_epi16 (res, 1); //double the result, don't care about saturation
     res_lo = _mm_mullo_epi16 (a, b);
@@ -4244,12 +4246,12 @@ _NEON2SSE_INLINE int16x8_t vqdmulhq_s16(int16x8_t a, int16x8_t b) // VQDMULH.S16
     return _mm_xor_si128 (res,  mask); //res saturated for 0x8000
 }
 
-int32x4_t vqdmulhq_s32(int32x4_t a, int32x4_t b); // VQDMULH.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vqdmulhq_s32(int32x4_t a, int32x4_t b); // VQDMULH.S32 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqdmulhq_s32(int32x4_t a, int32x4_t b), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     // no multiply high 32 bit SIMD in IA32, may be not optimal compared with a serial solution for the SSSE3 target
     __m128i ab, ba, mask, mul, mul1;
-    _NEON2SSE_ALIGN_16 uint32_t cmask32[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
+    _NEON2SSE_ALIGN_16 static const uint32_t cmask32[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
     ab = _mm_unpacklo_epi32 (a, b); //a0, b0, a1,b1
     ba = _mm_unpacklo_epi32 (b, a); //b0, a0, b1,a1
     mul = _MM_MUL_EPI32(ab, ba); //uses 1rst and 3rd data lanes, the multiplication gives 64 bit result
@@ -4268,19 +4270,19 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqdmulhq_s32(int32x4_t
 //********* Vector saturating rounding doubling multiply high ****************
 //****************************************************************************
 //If use _mm_mulhrs_xx functions  the result may differ from NEON one a little  due to different rounding rules and order
-int16x4_t vqrdmulh_s16(int16x4_t a,  int16x4_t b); // VQRDMULH.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vqrdmulh_s16(int16x4_t a,  int16x4_t b); // VQRDMULH.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vqrdmulh_s16(int16x4_t a,  int16x4_t b)
 {
     int16x4_t res64;
     return64(vqrdmulhq_s16(_pM128i(a), _pM128i(b)));
 }
 
-int32x2_t vqrdmulh_s32(int32x2_t a, int32x2_t b); // VQRDMULH.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqrdmulh_s32(int32x2_t a, int32x2_t b); // VQRDMULH.S32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqrdmulh_s32(int32x2_t a, int32x2_t b), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     //may be not optimal compared with a serial solution
     int32x2_t res64;
-    _NEON2SSE_ALIGN_16 uint32_t cmask32[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
+    _NEON2SSE_ALIGN_16 static const uint32_t cmask32[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
     __m128i res_sat, mask, mask1;
     int64x2_t mul;
     mul = vmull_s32(a,b);
@@ -4295,22 +4297,22 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqrdmulh_s32(int32x2_t
     return64(mul);
 }
 
-int16x8_t vqrdmulhq_s16(int16x8_t a, int16x8_t b); // VQRDMULH.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vqrdmulhq_s16(int16x8_t a, int16x8_t b); // VQRDMULH.S16 q0,q0,q0
 _NEON2SSE_INLINE int16x8_t vqrdmulhq_s16(int16x8_t a, int16x8_t b) // VQRDMULH.S16 q0,q0,q0
 {
     __m128i mask, res;
-    _NEON2SSE_ALIGN_16 uint16_t cmask[] = {0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000};
+    _NEON2SSE_ALIGN_16 static const uint16_t cmask[] = {0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000, 0x8000};
     res = _mm_mulhrs_epi16 (a, b);
     mask = _mm_cmpeq_epi16 (res, *(__m128i*)cmask);
     return _mm_xor_si128 (res,  mask); //res saturated for 0x8000
 }
 
-int32x4_t vqrdmulhq_s32(int32x4_t a, int32x4_t b); // VQRDMULH.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vqrdmulhq_s32(int32x4_t a, int32x4_t b); // VQRDMULH.S32 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqrdmulhq_s32(int32x4_t a, int32x4_t b), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     // no multiply high 32 bit SIMD in IA32, may be not optimal compared with a serial solution for the SSSE3 target
     __m128i ab, ba,  mask, mul, mul1, mask1;
-    _NEON2SSE_ALIGN_16 uint32_t cmask32[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
+    _NEON2SSE_ALIGN_16 static const uint32_t cmask32[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
     ab = _mm_unpacklo_epi32 (a, b); //a0, b0, a1,b1
     ba = _mm_unpacklo_epi32 (b, a); //b0, a0, b1,a1
     mul = _MM_MUL_EPI32(ab, ba); //uses 1rst and 3rd data lanes, the multiplication gives 64 bit result
@@ -4336,7 +4338,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqrdmulhq_s32(int32x4_t
 
 //*************Vector widening saturating doubling multiply accumulate (long saturating doubling multiply accumulate) *****
 //*************************************************************************************************************************
-int32x4_t vqdmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VQDMLAL.S16 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vqdmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VQDMLAL.S16 q0,d0,d0
 _NEON2SSE_INLINE int32x4_t vqdmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c) // VQDMLAL.S16 q0,d0,d0
 {
     //not optimal SIMD soulution, serial may be faster
@@ -4346,7 +4348,7 @@ _NEON2SSE_INLINE int32x4_t vqdmlal_s16(int32x4_t a, int16x4_t b, int16x4_t c) //
     return vqaddq_s32(res32, a); //saturation
 }
 
-int64x2_t vqdmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VQDMLAL.S32 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vqdmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VQDMLAL.S32 q0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmlal_s32(int64x2_t a, int32x2_t b, int32x2_t c),_NEON2SSE_REASON_SLOW_SERIAL)
 {
     __m128i res64;
@@ -4358,7 +4360,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmlal_s32(int64x2_t a
 //************************************************************************************
 //******************  Vector subtract ***********************************************
 //************************************************************************************
-int8x8_t vsub_s8(int8x8_t a, int8x8_t b); // VSUB.I8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vsub_s8(int8x8_t a, int8x8_t b); // VSUB.I8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vsub_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
@@ -4366,7 +4368,7 @@ _NEON2SSE_INLINE int8x8_t vsub_s8(int8x8_t a, int8x8_t b)
 }
 
 
-int16x4_t vsub_s16(int16x4_t a, int16x4_t b); // VSUB.I16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vsub_s16(int16x4_t a, int16x4_t b); // VSUB.I16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vsub_s16(int16x4_t a, int16x4_t b)
 {
     int16x4_t res64;
@@ -4374,7 +4376,7 @@ _NEON2SSE_INLINE int16x4_t vsub_s16(int16x4_t a, int16x4_t b)
 }
 
 
-int32x2_t vsub_s32(int32x2_t a, int32x2_t b); // VSUB.I32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vsub_s32(int32x2_t a, int32x2_t b); // VSUB.I32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vsub_s32(int32x2_t a, int32x2_t b)
 {
     int32x2_t res64;
@@ -4382,7 +4384,7 @@ _NEON2SSE_INLINE int32x2_t vsub_s32(int32x2_t a, int32x2_t b)
 }
 
 
-int64x1_t vsub_s64(int64x1_t a,  int64x1_t b); // VSUB.I64 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vsub_s64(int64x1_t a,  int64x1_t b); // VSUB.I64 d0,d0,d0
 _NEON2SSE_INLINE int64x1_t vsub_s64(int64x1_t a,  int64x1_t b)
 {
     int64x1_t res64;
@@ -4391,7 +4393,7 @@ _NEON2SSE_INLINE int64x1_t vsub_s64(int64x1_t a,  int64x1_t b)
 }
 
 
-float32x2_t vsub_f32(float32x2_t a, float32x2_t b); // VSUB.F32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vsub_f32(float32x2_t a, float32x2_t b); // VSUB.F32 d0,d0,d0
 _NEON2SSE_INLINE float32x2_t vsub_f32(float32x2_t a, float32x2_t b)
 {
     float32x2_t res;
@@ -4400,17 +4402,17 @@ _NEON2SSE_INLINE float32x2_t vsub_f32(float32x2_t a, float32x2_t b)
     return res;
 }
 
-uint8x8_t vsub_u8(uint8x8_t a, uint8x8_t b); // VSUB.I8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vsub_u8(uint8x8_t a, uint8x8_t b); // VSUB.I8 d0,d0,d0
 #define vsub_u8 vsub_s8
 
-uint16x4_t vsub_u16(uint16x4_t a, uint16x4_t b); // VSUB.I16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vsub_u16(uint16x4_t a, uint16x4_t b); // VSUB.I16 d0,d0,d0
 #define vsub_u16 vsub_s16
 
-uint32x2_t vsub_u32(uint32x2_t a, uint32x2_t b); // VSUB.I32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vsub_u32(uint32x2_t a, uint32x2_t b); // VSUB.I32 d0,d0,d0
 #define vsub_u32 vsub_s32
 
 
-uint64x1_t vsub_u64(uint64x1_t a,  uint64x1_t b); // VSUB.I64 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vsub_u64(uint64x1_t a,  uint64x1_t b); // VSUB.I64 d0,d0,d0
 _NEON2SSE_INLINE uint64x1_t vsub_u64(uint64x1_t a,  uint64x1_t b)
 {
     int64x1_t res64;
@@ -4419,37 +4421,37 @@ _NEON2SSE_INLINE uint64x1_t vsub_u64(uint64x1_t a,  uint64x1_t b)
 }
 
 
-int8x16_t   vsubq_s8(int8x16_t a, int8x16_t b); // VSUB.I8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t   vsubq_s8(int8x16_t a, int8x16_t b); // VSUB.I8 q0,q0,q0
 #define vsubq_s8 _mm_sub_epi8
 
-int16x8_t   vsubq_s16(int16x8_t a, int16x8_t b); // VSUB.I16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t   vsubq_s16(int16x8_t a, int16x8_t b); // VSUB.I16 q0,q0,q0
 #define vsubq_s16 _mm_sub_epi16
 
-int32x4_t   vsubq_s32(int32x4_t a, int32x4_t b); // VSUB.I32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t   vsubq_s32(int32x4_t a, int32x4_t b); // VSUB.I32 q0,q0,q0
 #define vsubq_s32 _mm_sub_epi32
 
-int64x2_t   vsubq_s64(int64x2_t a, int64x2_t b); // VSUB.I64 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t   vsubq_s64(int64x2_t a, int64x2_t b); // VSUB.I64 q0,q0,q0
 #define vsubq_s64 _mm_sub_epi64
 
-float32x4_t vsubq_f32(float32x4_t a, float32x4_t b); // VSUB.F32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vsubq_f32(float32x4_t a, float32x4_t b); // VSUB.F32 q0,q0,q0
 #define vsubq_f32 _mm_sub_ps
 
-uint8x16_t   vsubq_u8(uint8x16_t a, uint8x16_t b); // VSUB.I8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t   vsubq_u8(uint8x16_t a, uint8x16_t b); // VSUB.I8 q0,q0,q0
 #define vsubq_u8 _mm_sub_epi8
 
-uint16x8_t   vsubq_u16(uint16x8_t a, uint16x8_t b); // VSUB.I16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   vsubq_u16(uint16x8_t a, uint16x8_t b); // VSUB.I16 q0,q0,q0
 #define vsubq_u16 _mm_sub_epi16
 
-uint32x4_t   vsubq_u32(uint32x4_t a, uint32x4_t b); // VSUB.I32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t   vsubq_u32(uint32x4_t a, uint32x4_t b); // VSUB.I32 q0,q0,q0
 #define vsubq_u32 _mm_sub_epi32
 
-uint64x2_t   vsubq_u64(uint64x2_t a, uint64x2_t b); // VSUB.I64 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t   vsubq_u64(uint64x2_t a, uint64x2_t b); // VSUB.I64 q0,q0,q0
 #define vsubq_u64 _mm_sub_epi64
 
 //***************Vector long subtract: vsub -> Vr[i]:=Va[i]-Vb[i] ******************
 //***********************************************************************************
 //Va, Vb have equal lane sizes, result is a 128 bit vector of lanes that are twice the width.
-int16x8_t vsubl_s8(int8x8_t a, int8x8_t b); // VSUBL.S8 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vsubl_s8(int8x8_t a, int8x8_t b); // VSUBL.S8 q0,d0,d0
 _NEON2SSE_INLINE int16x8_t vsubl_s8(int8x8_t a, int8x8_t b) // VSUBL.S8 q0,d0,d0
 {
     __m128i a16, b16;
@@ -4458,7 +4460,7 @@ _NEON2SSE_INLINE int16x8_t vsubl_s8(int8x8_t a, int8x8_t b) // VSUBL.S8 q0,d0,d0
     return _mm_sub_epi16 (a16, b16);
 }
 
-int32x4_t vsubl_s16(int16x4_t a, int16x4_t b); // VSUBL.S16 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vsubl_s16(int16x4_t a, int16x4_t b); // VSUBL.S16 q0,d0,d0
 _NEON2SSE_INLINE int32x4_t vsubl_s16(int16x4_t a, int16x4_t b) // VSUBL.S16 q0,d0,d0
 {
     __m128i a32, b32;
@@ -4467,7 +4469,7 @@ _NEON2SSE_INLINE int32x4_t vsubl_s16(int16x4_t a, int16x4_t b) // VSUBL.S16 q0,d
     return _mm_sub_epi32 (a32, b32);
 }
 
-int64x2_t vsubl_s32(int32x2_t a, int32x2_t b); // VSUBL.S32 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vsubl_s32(int32x2_t a, int32x2_t b); // VSUBL.S32 q0,d0,d0
 _NEON2SSE_INLINE int64x2_t vsubl_s32(int32x2_t a, int32x2_t b) // VSUBL.S32 q0,d0,d0
 {
     //may be not optimal
@@ -4477,7 +4479,7 @@ _NEON2SSE_INLINE int64x2_t vsubl_s32(int32x2_t a, int32x2_t b) // VSUBL.S32 q0,d
     return _mm_sub_epi64 (a64, b64);
 }
 
-uint16x8_t vsubl_u8(uint8x8_t a, uint8x8_t b); // VSUBL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vsubl_u8(uint8x8_t a, uint8x8_t b); // VSUBL.U8 q0,d0,d0
 _NEON2SSE_INLINE uint16x8_t vsubl_u8(uint8x8_t a, uint8x8_t b) // VSUBL.U8 q0,d0,d0
 {
     __m128i a16, b16;
@@ -4486,7 +4488,7 @@ _NEON2SSE_INLINE uint16x8_t vsubl_u8(uint8x8_t a, uint8x8_t b) // VSUBL.U8 q0,d0
     return _mm_sub_epi16 (a16, b16);
 }
 
-uint32x4_t vsubl_u16(uint16x4_t a, uint16x4_t b); // VSUBL.s16 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vsubl_u16(uint16x4_t a, uint16x4_t b); // VSUBL.s16 q0,d0,d0
 _NEON2SSE_INLINE uint32x4_t vsubl_u16(uint16x4_t a, uint16x4_t b) // VSUBL.s16 q0,d0,d0
 {
     __m128i a32, b32;
@@ -4495,7 +4497,7 @@ _NEON2SSE_INLINE uint32x4_t vsubl_u16(uint16x4_t a, uint16x4_t b) // VSUBL.s16 q
     return _mm_sub_epi32 (a32, b32);
 }
 
-uint64x2_t vsubl_u32(uint32x2_t a, uint32x2_t b); // VSUBL.U32 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vsubl_u32(uint32x2_t a, uint32x2_t b); // VSUBL.U32 q0,d0,d0
 _NEON2SSE_INLINE uint64x2_t vsubl_u32(uint32x2_t a, uint32x2_t b) // VSUBL.U32 q0,d0,d0
 {
     //may be not optimal
@@ -4507,7 +4509,7 @@ _NEON2SSE_INLINE uint64x2_t vsubl_u32(uint32x2_t a, uint32x2_t b) // VSUBL.U32 q
 
 //***************** Vector wide subtract: vsub -> Vr[i]:=Va[i]-Vb[i] **********************************
 //*****************************************************************************************************
-int16x8_t vsubw_s8(int16x8_t a, int8x8_t b); // VSUBW.S8 q0,q0,d0
+_NEON2SSESTORAGE int16x8_t vsubw_s8(int16x8_t a, int8x8_t b); // VSUBW.S8 q0,q0,d0
 _NEON2SSE_INLINE int16x8_t vsubw_s8(int16x8_t a, int8x8_t b) // VSUBW.S8 q0,q0,d0
 {
     __m128i b16;
@@ -4515,7 +4517,7 @@ _NEON2SSE_INLINE int16x8_t vsubw_s8(int16x8_t a, int8x8_t b) // VSUBW.S8 q0,q0,d
     return _mm_sub_epi16 (a, b16);
 }
 
-int32x4_t vsubw_s16(int32x4_t a, int16x4_t b); // VSUBW.S16 q0,q0,d0
+_NEON2SSESTORAGE int32x4_t vsubw_s16(int32x4_t a, int16x4_t b); // VSUBW.S16 q0,q0,d0
 _NEON2SSE_INLINE int32x4_t vsubw_s16(int32x4_t a, int16x4_t b) // VSUBW.S16 q0,q0,d0
 {
     __m128i b32;
@@ -4523,7 +4525,7 @@ _NEON2SSE_INLINE int32x4_t vsubw_s16(int32x4_t a, int16x4_t b) // VSUBW.S16 q0,q
     return _mm_sub_epi32 (a, b32);
 }
 
-int64x2_t vsubw_s32(int64x2_t a, int32x2_t b); // VSUBW.S32 q0,q0,d0
+_NEON2SSESTORAGE int64x2_t vsubw_s32(int64x2_t a, int32x2_t b); // VSUBW.S32 q0,q0,d0
 _NEON2SSE_INLINE int64x2_t vsubw_s32(int64x2_t a, int32x2_t b) // VSUBW.S32 q0,q0,d0
 {
     __m128i b64;
@@ -4531,7 +4533,7 @@ _NEON2SSE_INLINE int64x2_t vsubw_s32(int64x2_t a, int32x2_t b) // VSUBW.S32 q0,q
     return _mm_sub_epi64 (a, b64);
 }
 
-uint16x8_t vsubw_u8(uint16x8_t a, uint8x8_t b); // VSUBW.U8 q0,q0,d0
+_NEON2SSESTORAGE uint16x8_t vsubw_u8(uint16x8_t a, uint8x8_t b); // VSUBW.U8 q0,q0,d0
 _NEON2SSE_INLINE uint16x8_t vsubw_u8(uint16x8_t a, uint8x8_t b) // VSUBW.U8 q0,q0,d0
 {
     __m128i b16;
@@ -4539,7 +4541,7 @@ _NEON2SSE_INLINE uint16x8_t vsubw_u8(uint16x8_t a, uint8x8_t b) // VSUBW.U8 q0,q
     return _mm_sub_epi16 (a, b16);
 }
 
-uint32x4_t vsubw_u16(uint32x4_t a, uint16x4_t b); // VSUBW.s16 q0,q0,d0
+_NEON2SSESTORAGE uint32x4_t vsubw_u16(uint32x4_t a, uint16x4_t b); // VSUBW.s16 q0,q0,d0
 _NEON2SSE_INLINE uint32x4_t vsubw_u16(uint32x4_t a, uint16x4_t b) // VSUBW.s16 q0,q0,d0
 {
     __m128i b32;
@@ -4547,7 +4549,7 @@ _NEON2SSE_INLINE uint32x4_t vsubw_u16(uint32x4_t a, uint16x4_t b) // VSUBW.s16 q
     return _mm_sub_epi32 (a, b32);
 }
 
-uint64x2_t vsubw_u32(uint64x2_t a, uint32x2_t b); // VSUBW.U32 q0,q0,d0
+_NEON2SSESTORAGE uint64x2_t vsubw_u32(uint64x2_t a, uint32x2_t b); // VSUBW.U32 q0,q0,d0
 _NEON2SSE_INLINE uint64x2_t vsubw_u32(uint64x2_t a, uint32x2_t b) // VSUBW.U32 q0,q0,d0
 {
     __m128i b64;
@@ -4557,7 +4559,7 @@ _NEON2SSE_INLINE uint64x2_t vsubw_u32(uint64x2_t a, uint32x2_t b) // VSUBW.U32 q
 
 //************************Vector saturating subtract *********************************
 //*************************************************************************************
-int8x8_t vqsub_s8(int8x8_t a, int8x8_t b); // VQSUB.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vqsub_s8(int8x8_t a, int8x8_t b); // VQSUB.S8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vqsub_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
@@ -4565,7 +4567,7 @@ _NEON2SSE_INLINE int8x8_t vqsub_s8(int8x8_t a, int8x8_t b)
 }
 
 
-int16x4_t vqsub_s16(int16x4_t a, int16x4_t b); // VQSUB.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vqsub_s16(int16x4_t a, int16x4_t b); // VQSUB.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vqsub_s16(int16x4_t a, int16x4_t b)
 {
     int16x4_t res64;
@@ -4573,7 +4575,7 @@ _NEON2SSE_INLINE int16x4_t vqsub_s16(int16x4_t a, int16x4_t b)
 }
 
 
-int32x2_t vqsub_s32(int32x2_t a,  int32x2_t b); // VQSUB.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqsub_s32(int32x2_t a,  int32x2_t b); // VQSUB.S32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vqsub_s32(int32x2_t a,  int32x2_t b)
 {
     int32x2_t res64;
@@ -4581,7 +4583,7 @@ _NEON2SSE_INLINE int32x2_t vqsub_s32(int32x2_t a,  int32x2_t b)
 }
 
 
-int64x1_t vqsub_s64(int64x1_t a, int64x1_t b); // VQSUB.S64 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vqsub_s64(int64x1_t a, int64x1_t b); // VQSUB.S64 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vqsub_s64(int64x1_t a, int64x1_t b), _NEON2SSE_REASON_SLOW_SERIAL) //no optimal SIMD soulution
 {
     uint64x1_t res;
@@ -4597,7 +4599,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vqsub_s64(int64x1_t a,
     return res;
 }
 
-uint8x8_t vqsub_u8(uint8x8_t a, uint8x8_t b); // VQSUB.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vqsub_u8(uint8x8_t a, uint8x8_t b); // VQSUB.U8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vqsub_u8(uint8x8_t a, uint8x8_t b)
 {
     uint8x8_t res64;
@@ -4605,7 +4607,7 @@ _NEON2SSE_INLINE uint8x8_t vqsub_u8(uint8x8_t a, uint8x8_t b)
 }
 
 
-uint16x4_t vqsub_u16(uint16x4_t a, uint16x4_t b); // VQSUB.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vqsub_u16(uint16x4_t a, uint16x4_t b); // VQSUB.s16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vqsub_u16(uint16x4_t a, uint16x4_t b)
 {
     uint16x4_t res64;
@@ -4613,7 +4615,7 @@ _NEON2SSE_INLINE uint16x4_t vqsub_u16(uint16x4_t a, uint16x4_t b)
 }
 
 
-uint32x2_t vqsub_u32(uint32x2_t a,  uint32x2_t b); // VQSUB.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vqsub_u32(uint32x2_t a,  uint32x2_t b); // VQSUB.U32 d0,d0,d0
 _NEON2SSE_INLINE uint32x2_t vqsub_u32(uint32x2_t a,  uint32x2_t b)
 {
     uint32x2_t res64;
@@ -4621,7 +4623,7 @@ _NEON2SSE_INLINE uint32x2_t vqsub_u32(uint32x2_t a,  uint32x2_t b)
 }
 
 
-uint64x1_t vqsub_u64(uint64x1_t a, uint64x1_t b); // VQSUB.U64 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vqsub_u64(uint64x1_t a, uint64x1_t b); // VQSUB.U64 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vqsub_u64(uint64x1_t a, uint64x1_t b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     uint64x1_t res;
@@ -4636,13 +4638,13 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vqsub_u64(uint64x1_t a
     return res;
 }
 
-int8x16_t   vqsubq_s8(int8x16_t a, int8x16_t b); // VQSUB.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t   vqsubq_s8(int8x16_t a, int8x16_t b); // VQSUB.S8 q0,q0,q0
 #define vqsubq_s8 _mm_subs_epi8
 
-int16x8_t   vqsubq_s16(int16x8_t a, int16x8_t b); // VQSUB.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t   vqsubq_s16(int16x8_t a, int16x8_t b); // VQSUB.S16 q0,q0,q0
 #define vqsubq_s16 _mm_subs_epi16
 
-int32x4_t vqsubq_s32(int32x4_t a, int32x4_t b); // VQSUB.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vqsubq_s32(int32x4_t a, int32x4_t b); // VQSUB.S32 q0,q0,q0
 _NEON2SSE_INLINE int32x4_t vqsubq_s32(int32x4_t a, int32x4_t b)
 {
     //no corresponding x86 SIMD soulution, special tricks are necessary. The overflow is possible only if a and b have opposite signs and sub has opposite sign to a
@@ -4660,7 +4662,7 @@ _NEON2SSE_INLINE int32x4_t vqsubq_s32(int32x4_t a, int32x4_t b)
     return _mm_or_si128(res, res_sat);
 }
 
-int64x2_t vqsubq_s64(int64x2_t a, int64x2_t b); // VQSUB.S64 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vqsubq_s64(int64x2_t a, int64x2_t b); // VQSUB.S64 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqsubq_s64(int64x2_t a, int64x2_t b), _NEON2SSE_REASON_SLOW_SERIAL) //no optimal SIMD soulution
 {
     _NEON2SSE_ALIGN_16 int64_t atmp[2], btmp[2];
@@ -4678,13 +4680,13 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqsubq_s64(int64x2_t a,
     return _mm_load_si128((__m128i*)res);
 }
 
-uint8x16_t   vqsubq_u8(uint8x16_t a, uint8x16_t b); // VQSUB.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t   vqsubq_u8(uint8x16_t a, uint8x16_t b); // VQSUB.U8 q0,q0,q0
 #define vqsubq_u8 _mm_subs_epu8
 
-uint16x8_t   vqsubq_u16(uint16x8_t a, uint16x8_t b); // VQSUB.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   vqsubq_u16(uint16x8_t a, uint16x8_t b); // VQSUB.s16 q0,q0,q0
 #define vqsubq_u16 _mm_subs_epu16
 
-uint32x4_t vqsubq_u32(uint32x4_t a, uint32x4_t b); // VQSUB.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vqsubq_u32(uint32x4_t a, uint32x4_t b); // VQSUB.U32 q0,q0,q0
 _NEON2SSE_INLINE uint32x4_t vqsubq_u32(uint32x4_t a, uint32x4_t b) // VQSUB.U32 q0,q0,q0
 {
     __m128i min, mask, sub;
@@ -4720,7 +4722,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vqsubq_u64(uint64x2_t
 
 //**********Vector halving subtract Vr[i]:=(Va[i]-Vb[i])>>1  ******************************************************
 //****************************************************************
-int8x8_t vhsub_s8(int8x8_t a, int8x8_t b); // VHSUB.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vhsub_s8(int8x8_t a, int8x8_t b); // VHSUB.S8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vhsub_s8(int8x8_t a, int8x8_t b) // VHSUB.S8 d0,d0,d0
 {
     //no 8 bit shift available, internal overflow is possible, so let's go to 16 bit,
@@ -4734,7 +4736,7 @@ _NEON2SSE_INLINE int8x8_t vhsub_s8(int8x8_t a, int8x8_t b) // VHSUB.S8 d0,d0,d0
     return64(r16);
 }
 
-int16x4_t vhsub_s16(int16x4_t a,  int16x4_t b); // VHSUB.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vhsub_s16(int16x4_t a,  int16x4_t b); // VHSUB.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vhsub_s16(int16x4_t a,  int16x4_t b)
 {
     int16x4_t res64;
@@ -4743,7 +4745,7 @@ _NEON2SSE_INLINE int16x4_t vhsub_s16(int16x4_t a,  int16x4_t b)
 
 
 
-int32x2_t vhsub_s32(int32x2_t a,  int32x2_t b); // VHSUB.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vhsub_s32(int32x2_t a,  int32x2_t b); // VHSUB.S32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vhsub_s32(int32x2_t a,  int32x2_t b)
 {
     int32x2_t res64;
@@ -4751,28 +4753,28 @@ _NEON2SSE_INLINE int32x2_t vhsub_s32(int32x2_t a,  int32x2_t b)
 }
 
 
-uint8x8_t vhsub_u8(uint8x8_t a,  uint8x8_t b); // VHSUB.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vhsub_u8(uint8x8_t a,  uint8x8_t b); // VHSUB.U8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vhsub_u8(uint8x8_t a,  uint8x8_t b)
 {
     uint8x8_t res64;
     return64(vhsubq_u8(_pM128i(a), _pM128i(b)));
 }
 
-uint16x4_t vhsub_u16(uint16x4_t a,  uint16x4_t b); // VHSUB.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vhsub_u16(uint16x4_t a,  uint16x4_t b); // VHSUB.s16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vhsub_u16(uint16x4_t a,  uint16x4_t b)
 {
     uint16x4_t res64;
     return64(vhsubq_u16(_pM128i(a), _pM128i(b)));
 }
 
-uint32x2_t vhsub_u32(uint32x2_t a,  uint32x2_t b); // VHSUB.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vhsub_u32(uint32x2_t a,  uint32x2_t b); // VHSUB.U32 d0,d0,d0
 _NEON2SSE_INLINE uint32x2_t vhsub_u32(uint32x2_t a,  uint32x2_t b)
 {
     uint32x2_t res64;
     return64(vhsubq_u32(_pM128i(a), _pM128i(b)));
 }
 
-int8x16_t vhsubq_s8(int8x16_t a, int8x16_t b); // VHSUB.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vhsubq_s8(int8x16_t a, int8x16_t b); // VHSUB.S8 q0,q0,q0
 _NEON2SSE_INLINE int8x16_t vhsubq_s8(int8x16_t a, int8x16_t b) // VHSUB.S8 q0,q0,q0
 {
     // //need to deal with the possibility of internal overflow
@@ -4783,7 +4785,7 @@ _NEON2SSE_INLINE int8x16_t vhsubq_s8(int8x16_t a, int8x16_t b) // VHSUB.S8 q0,q0
     return vhsubq_u8(au,bu);
 }
 
-int16x8_t vhsubq_s16(int16x8_t a, int16x8_t b); // VHSUB.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vhsubq_s16(int16x8_t a, int16x8_t b); // VHSUB.S16 q0,q0,q0
 _NEON2SSE_INLINE int16x8_t vhsubq_s16(int16x8_t a, int16x8_t b) // VHSUB.S16 q0,q0,q0
 {
     //need to deal with the possibility of internal overflow
@@ -4794,7 +4796,7 @@ _NEON2SSE_INLINE int16x8_t vhsubq_s16(int16x8_t a, int16x8_t b) // VHSUB.S16 q0,
     return vhsubq_u16(au,bu);
 }
 
-int32x4_t vhsubq_s32(int32x4_t a, int32x4_t b); // VHSUB.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vhsubq_s32(int32x4_t a, int32x4_t b); // VHSUB.S32 q0,q0,q0
 _NEON2SSE_INLINE int32x4_t vhsubq_s32(int32x4_t a, int32x4_t b) // VHSUB.S32 q0,q0,q0
 {
     //need to deal with the possibility of internal overflow
@@ -4808,7 +4810,7 @@ _NEON2SSE_INLINE int32x4_t vhsubq_s32(int32x4_t a, int32x4_t b) // VHSUB.S32 q0,
     return _mm_sub_epi32(r,b_1);
 }
 
-uint8x16_t vhsubq_u8(uint8x16_t a, uint8x16_t b); // VHSUB.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vhsubq_u8(uint8x16_t a, uint8x16_t b); // VHSUB.U8 q0,q0,q0
 _NEON2SSE_INLINE uint8x16_t vhsubq_u8(uint8x16_t a, uint8x16_t b) // VHSUB.U8 q0,q0,q0
 {
     __m128i avg;
@@ -4816,7 +4818,7 @@ _NEON2SSE_INLINE uint8x16_t vhsubq_u8(uint8x16_t a, uint8x16_t b) // VHSUB.U8 q0
     return _mm_sub_epi8(a, avg);
 }
 
-uint16x8_t vhsubq_u16(uint16x8_t a, uint16x8_t b); // VHSUB.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vhsubq_u16(uint16x8_t a, uint16x8_t b); // VHSUB.s16 q0,q0,q0
 _NEON2SSE_INLINE uint16x8_t vhsubq_u16(uint16x8_t a, uint16x8_t b) // VHSUB.s16 q0,q0,q0
 {
     __m128i avg;
@@ -4824,7 +4826,7 @@ _NEON2SSE_INLINE uint16x8_t vhsubq_u16(uint16x8_t a, uint16x8_t b) // VHSUB.s16
     return _mm_sub_epi16(a, avg);
 }
 
-uint32x4_t vhsubq_u32(uint32x4_t a, uint32x4_t b); // VHSUB.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vhsubq_u32(uint32x4_t a, uint32x4_t b); // VHSUB.U32 q0,q0,q0
 _NEON2SSE_INLINE uint32x4_t vhsubq_u32(uint32x4_t a, uint32x4_t b) // VHSUB.U32 q0,q0,q0
 {
     //need to deal with the possibility of internal overflow
@@ -4840,7 +4842,7 @@ _NEON2SSE_INLINE uint32x4_t vhsubq_u32(uint32x4_t a, uint32x4_t b) // VHSUB.U32
 
 //******* Vector subtract high half (truncated) ** ************
 //************************************************************
-int8x8_t vsubhn_s16(int16x8_t a, int16x8_t b); // VSUBHN.I16 d0,q0,q0
+_NEON2SSESTORAGE int8x8_t vsubhn_s16(int16x8_t a, int16x8_t b); // VSUBHN.I16 d0,q0,q0
 _NEON2SSE_INLINE int8x8_t vsubhn_s16(int16x8_t a, int16x8_t b) // VSUBHN.I16 d0,q0,q0
 {
     int8x8_t res64;
@@ -4851,7 +4853,7 @@ _NEON2SSE_INLINE int8x8_t vsubhn_s16(int16x8_t a, int16x8_t b) // VSUBHN.I16 d0,
     return64(sum8);
 }
 
-int16x4_t vsubhn_s32(int32x4_t a, int32x4_t b); // VSUBHN.I32 d0,q0,q0
+_NEON2SSESTORAGE int16x4_t vsubhn_s32(int32x4_t a, int32x4_t b); // VSUBHN.I32 d0,q0,q0
 _NEON2SSE_INLINE int16x4_t vsubhn_s32(int32x4_t a, int32x4_t b) // VSUBHN.I32 d0,q0,q0
 {
     int16x4_t res64;
@@ -4862,7 +4864,7 @@ _NEON2SSE_INLINE int16x4_t vsubhn_s32(int32x4_t a, int32x4_t b) // VSUBHN.I32 d0
     return64(sum16);
 }
 
-int32x2_t vsubhn_s64(int64x2_t a, int64x2_t b); // VSUBHN.I64 d0,q0,q0
+_NEON2SSESTORAGE int32x2_t vsubhn_s64(int64x2_t a, int64x2_t b); // VSUBHN.I64 d0,q0,q0
 _NEON2SSE_INLINE int32x2_t vsubhn_s64(int64x2_t a, int64x2_t b)
 {
     int32x2_t res64;
@@ -4872,7 +4874,7 @@ _NEON2SSE_INLINE int32x2_t vsubhn_s64(int64x2_t a, int64x2_t b)
     return64(sub);
 }
 
-uint8x8_t vsubhn_u16(uint16x8_t a, uint16x8_t b); // VSUBHN.I16 d0,q0,q0
+_NEON2SSESTORAGE uint8x8_t vsubhn_u16(uint16x8_t a, uint16x8_t b); // VSUBHN.I16 d0,q0,q0
 _NEON2SSE_INLINE uint8x8_t vsubhn_u16(uint16x8_t a, uint16x8_t b) // VSUBHN.I16 d0,q0,q0
 {
     uint8x8_t res64;
@@ -4883,27 +4885,27 @@ _NEON2SSE_INLINE uint8x8_t vsubhn_u16(uint16x8_t a, uint16x8_t b) // VSUBHN.I16
     return64(sum8);
 }
 
-uint16x4_t vsubhn_u32(uint32x4_t a, uint32x4_t b); // VSUBHN.I32 d0,q0,q0
+_NEON2SSESTORAGE uint16x4_t vsubhn_u32(uint32x4_t a, uint32x4_t b); // VSUBHN.I32 d0,q0,q0
 _NEON2SSE_INLINE uint16x4_t vsubhn_u32(uint32x4_t a, uint32x4_t b) // VSUBHN.I32 d0,q0,q0
 {
     uint16x4_t res64;
      __m128i sum, sum16;
     sum = _mm_sub_epi32 (a, b);
     sum16 = _mm_srli_epi32 (sum, 16);
-     #ifdef USE_SSE4
-        sum16 =  _MM_PACKUS1_EPI32(sum16);
-    #else
-        sum16  = _mm_shuffle_epi8 (sum16, *(__m128i*) mask8_32_even_odd); //go to 16 bits 
-    #endif
+#ifdef USE_SSE4
+    sum16 =  _MM_PACKUS1_EPI32(sum16);
+#else
+    sum16  = _mm_shuffle_epi8 (sum16, *(__m128i*) mask8_32_even_odd); //go to 16 bits
+#endif
     return64(sum16);
 }
 
-uint32x2_t vsubhn_u64(uint64x2_t a, uint64x2_t b); // VSUBHN.I64 d0,q0,q0
+_NEON2SSESTORAGE uint32x2_t vsubhn_u64(uint64x2_t a, uint64x2_t b); // VSUBHN.I64 d0,q0,q0
 #define vsubhn_u64 vsubhn_s64
 
 //************ Vector rounding subtract high half *********************
 //*********************************************************************
-int8x8_t vrsubhn_s16(int16x8_t a, int16x8_t b); // VRSUBHN.I16 d0,q0,q0
+_NEON2SSESTORAGE int8x8_t vrsubhn_s16(int16x8_t a, int16x8_t b); // VRSUBHN.I16 d0,q0,q0
 _NEON2SSE_INLINE int8x8_t vrsubhn_s16(int16x8_t a, int16x8_t b) // VRSUBHN.I16 d0,q0,q0
 {
     int8x8_t res64;
@@ -4917,7 +4919,7 @@ _NEON2SSE_INLINE int8x8_t vrsubhn_s16(int16x8_t a, int16x8_t b) // VRSUBHN.I16 d
     return64(sub);
 }
 
-int16x4_t vrsubhn_s32(int32x4_t a, int32x4_t b); // VRSUBHN.I32 d0,q0,q0
+_NEON2SSESTORAGE int16x4_t vrsubhn_s32(int32x4_t a, int32x4_t b); // VRSUBHN.I32 d0,q0,q0
 _NEON2SSE_INLINE int16x4_t vrsubhn_s32(int32x4_t a, int32x4_t b) // VRSUBHN.I32 d0,q0,q0
 {
     //SIMD may be not optimal, serial may be faster
@@ -4932,7 +4934,7 @@ _NEON2SSE_INLINE int16x4_t vrsubhn_s32(int32x4_t a, int32x4_t b) // VRSUBHN.I32
     return64(sub);
 }
 
-int32x2_t vrsubhn_s64(int64x2_t a, int64x2_t b); // VRSUBHN.I64 d0,q0,q0
+_NEON2SSESTORAGE int32x2_t vrsubhn_s64(int64x2_t a, int64x2_t b); // VRSUBHN.I64 d0,q0,q0
 _NEON2SSE_INLINE int32x2_t vrsubhn_s64(int64x2_t a, int64x2_t b)
 {
     //SIMD may be not optimal, serial may be faster
@@ -4946,7 +4948,7 @@ _NEON2SSE_INLINE int32x2_t vrsubhn_s64(int64x2_t a, int64x2_t b)
     return64(sub);
 }
 
-uint8x8_t vrsubhn_u16(uint16x8_t a, uint16x8_t b); // VRSUBHN.I16 d0,q0,q0
+_NEON2SSESTORAGE uint8x8_t vrsubhn_u16(uint16x8_t a, uint16x8_t b); // VRSUBHN.I16 d0,q0,q0
 _NEON2SSE_INLINE uint8x8_t vrsubhn_u16(uint16x8_t a, uint16x8_t b) // VRSUBHN.I16 d0,q0,q0
 {
     uint8x8_t res64;
@@ -4960,7 +4962,7 @@ _NEON2SSE_INLINE uint8x8_t vrsubhn_u16(uint16x8_t a, uint16x8_t b) // VRSUBHN.I1
     return64(sub);
 }
 
-uint16x4_t vrsubhn_u32(uint32x4_t a, uint32x4_t b); // VRSUBHN.I32 d0,q0,q0
+_NEON2SSESTORAGE uint16x4_t vrsubhn_u32(uint32x4_t a, uint32x4_t b); // VRSUBHN.I32 d0,q0,q0
 _NEON2SSE_INLINE uint16x4_t vrsubhn_u32(uint32x4_t a, uint32x4_t b) // VRSUBHN.I32 d0,q0,q0
 {
     //SIMD may be not optimal, serial may be faster
@@ -4971,26 +4973,26 @@ _NEON2SSE_INLINE uint16x4_t vrsubhn_u32(uint32x4_t a, uint32x4_t b) // VRSUBHN.I
     mask1 = _mm_srli_epi32(mask1,31); //get  15-th bit 1 or zero
     sub = _mm_srai_epi32 (sub, 16); //get high half
     sub = _mm_add_epi32 (sub, mask1); //actual rounding
-    #ifdef USE_SSE4
-        sub =  _MM_PACKUS1_EPI32 (sub);
-    #else
-        sub = _mm_shuffle_epi8 (sub, *(__m128i*) mask8_32_even_odd); //go to 16 bits 
-    #endif
+#ifdef USE_SSE4
+    sub =  _MM_PACKUS1_EPI32 (sub);
+#else
+    sub = _mm_shuffle_epi8 (sub, *(__m128i*) mask8_32_even_odd); //go to 16 bits
+#endif
     return64(sub);
 }
 
-uint32x2_t vrsubhn_u64(uint64x2_t a, uint64x2_t b); // VRSUBHN.I64 d0,q0,q0
+_NEON2SSESTORAGE uint32x2_t vrsubhn_u64(uint64x2_t a, uint64x2_t b); // VRSUBHN.I64 d0,q0,q0
 #define vrsubhn_u64 vrsubhn_s64
 
 //*********** Vector saturating doubling multiply subtract long ********************
 //************************************************************************************
-int32x4_t vqdmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VQDMLSL.S16 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vqdmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VQDMLSL.S16 q0,d0,d0
 _NEON2SSE_INLINE int32x4_t vqdmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c)
 {
     //not optimal SIMD soulution, serial may be faster
     __m128i res32, mask;
     int32x4_t res;
-    _NEON2SSE_ALIGN_16 uint32_t cmask[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
+    _NEON2SSE_ALIGN_16 static const uint32_t cmask[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
     res = vmull_s16(b,  c);
     res32 = _mm_slli_epi32 (res, 1); //double the result, saturation not considered
     mask = _mm_cmpeq_epi32 (res32, *(__m128i*)cmask);
@@ -4998,12 +5000,12 @@ _NEON2SSE_INLINE int32x4_t vqdmlsl_s16(int32x4_t a, int16x4_t b, int16x4_t c)
     return vqsubq_s32(a, res32); //saturation
 }
 
-int64x2_t vqdmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VQDMLSL.S32 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vqdmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VQDMLSL.S32 q0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmlsl_s32(int64x2_t a, int32x2_t b, int32x2_t c), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     __m128i res64, mask;
     int64x2_t res;
-    _NEON2SSE_ALIGN_16 uint64_t cmask[] = {0x8000000000000000, 0x8000000000000000};
+    _NEON2SSE_ALIGN_16 static const uint64_t cmask[] = {0x8000000000000000, 0x8000000000000000};
     res = vmull_s32(b,  c);
     res64 = _mm_slli_epi64 (res, 1); //double the result, saturation not considered
     mask = _MM_CMPEQ_EPI64 (res64, *(__m128i*)cmask);
@@ -5014,7 +5016,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmlsl_s32(int64x2_t a
 //******************  COMPARISON ***************************************
 //******************* Vector compare equal *************************************
 //****************************************************************************
-uint8x8_t vceq_s8(int8x8_t a, int8x8_t b); // VCEQ.I8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vceq_s8(int8x8_t a, int8x8_t b); // VCEQ.I8 d0, d0, d0
 _NEON2SSE_INLINE int8x8_t vceq_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
@@ -5022,7 +5024,7 @@ _NEON2SSE_INLINE int8x8_t vceq_s8(int8x8_t a, int8x8_t b)
 }
 
 
-uint16x4_t vceq_s16(int16x4_t a, int16x4_t b); // VCEQ.I16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vceq_s16(int16x4_t a, int16x4_t b); // VCEQ.I16 d0, d0, d0
 _NEON2SSE_INLINE int16x4_t vceq_s16(int16x4_t a, int16x4_t b)
 {
     int16x4_t res64;
@@ -5030,7 +5032,7 @@ _NEON2SSE_INLINE int16x4_t vceq_s16(int16x4_t a, int16x4_t b)
 }
 
 
-uint32x2_t vceq_s32(int32x2_t a, int32x2_t b); // VCEQ.I32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vceq_s32(int32x2_t a, int32x2_t b); // VCEQ.I32 d0, d0, d0
 _NEON2SSE_INLINE int32x2_t vceq_s32(int32x2_t a, int32x2_t b)
 {
     int32x2_t res64;
@@ -5038,7 +5040,7 @@ _NEON2SSE_INLINE int32x2_t vceq_s32(int32x2_t a, int32x2_t b)
 }
 
 
-uint32x2_t vceq_f32(float32x2_t a, float32x2_t b); // VCEQ.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vceq_f32(float32x2_t a, float32x2_t b); // VCEQ.F32 d0, d0, d0
 _NEON2SSE_INLINE uint32x2_t vceq_f32(float32x2_t a, float32x2_t b)
 {
     uint32x2_t res64;
@@ -5047,7 +5049,7 @@ _NEON2SSE_INLINE uint32x2_t vceq_f32(float32x2_t a, float32x2_t b)
     return64f(res);
 }
 
-uint8x8_t vceq_u8(uint8x8_t a, uint8x8_t b); // VCEQ.I8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vceq_u8(uint8x8_t a, uint8x8_t b); // VCEQ.I8 d0, d0, d0
 _NEON2SSE_INLINE uint8x8_t vceq_u8(uint8x8_t a, uint8x8_t b)
 {
     uint8x8_t res64;
@@ -5055,7 +5057,7 @@ _NEON2SSE_INLINE uint8x8_t vceq_u8(uint8x8_t a, uint8x8_t b)
 }
 
 
-uint16x4_t vceq_u16(uint16x4_t a, uint16x4_t b); // VCEQ.I16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vceq_u16(uint16x4_t a, uint16x4_t b); // VCEQ.I16 d0, d0, d0
 _NEON2SSE_INLINE uint16x4_t vceq_u16(uint16x4_t a, uint16x4_t b)
 {
     uint16x4_t res64;
@@ -5063,7 +5065,7 @@ _NEON2SSE_INLINE uint16x4_t vceq_u16(uint16x4_t a, uint16x4_t b)
 }
 
 
-uint32x2_t vceq_u32(uint32x2_t a, uint32x2_t b); // VCEQ.I32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vceq_u32(uint32x2_t a, uint32x2_t b); // VCEQ.I32 d0, d0, d0
 _NEON2SSE_INLINE uint32x2_t vceq_u32(uint32x2_t a, uint32x2_t b)
 {
     uint32x2_t res64;
@@ -5071,20 +5073,20 @@ _NEON2SSE_INLINE uint32x2_t vceq_u32(uint32x2_t a, uint32x2_t b)
 }
 
 
-uint8x8_t   vceq_p8(poly8x8_t a, poly8x8_t b); // VCEQ.I8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t   vceq_p8(poly8x8_t a, poly8x8_t b); // VCEQ.I8 d0, d0, d0
 #define vceq_p8 vceq_u8
 
 
-uint8x16_t   vceqq_s8(int8x16_t a, int8x16_t b); // VCEQ.I8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t   vceqq_s8(int8x16_t a, int8x16_t b); // VCEQ.I8 q0, q0, q0
 #define vceqq_s8 _mm_cmpeq_epi8
 
-uint16x8_t   vceqq_s16(int16x8_t a, int16x8_t b); // VCEQ.I16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t   vceqq_s16(int16x8_t a, int16x8_t b); // VCEQ.I16 q0, q0, q0
 #define vceqq_s16 _mm_cmpeq_epi16
 
-uint32x4_t   vceqq_s32(int32x4_t a, int32x4_t b); // VCEQ.I32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t   vceqq_s32(int32x4_t a, int32x4_t b); // VCEQ.I32 q0, q0, q0
 #define vceqq_s32 _mm_cmpeq_epi32
 
-uint32x4_t vceqq_f32(float32x4_t a, float32x4_t b); // VCEQ.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vceqq_f32(float32x4_t a, float32x4_t b); // VCEQ.F32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vceqq_f32(float32x4_t a, float32x4_t b)
 {
     __m128 res;
@@ -5092,16 +5094,16 @@ _NEON2SSE_INLINE uint32x4_t vceqq_f32(float32x4_t a, float32x4_t b)
     return _M128i(res);
 }
 
-uint8x16_t   vceqq_u8(uint8x16_t a, uint8x16_t b); // VCEQ.I8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t   vceqq_u8(uint8x16_t a, uint8x16_t b); // VCEQ.I8 q0, q0, q0
 #define vceqq_u8 _mm_cmpeq_epi8
 
-uint16x8_t   vceqq_u16(uint16x8_t a, uint16x8_t b); // VCEQ.I16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t   vceqq_u16(uint16x8_t a, uint16x8_t b); // VCEQ.I16 q0, q0, q0
 #define vceqq_u16 _mm_cmpeq_epi16
 
-uint32x4_t   vceqq_u32(uint32x4_t a, uint32x4_t b); // VCEQ.I32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t   vceqq_u32(uint32x4_t a, uint32x4_t b); // VCEQ.I32 q0, q0, q0
 #define vceqq_u32 _mm_cmpeq_epi32
 
-uint8x16_t   vceqq_p8(poly8x16_t a, poly8x16_t b); // VCEQ.I8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t   vceqq_p8(poly8x16_t a, poly8x16_t b); // VCEQ.I8 q0, q0, q0
 #define vceqq_p8 _mm_cmpeq_epi8
 
 //******************Vector compare greater-than or equal*************************
@@ -5109,7 +5111,7 @@ uint8x16_t   vceqq_p8(poly8x16_t a, poly8x16_t b); // VCEQ.I8 q0, q0, q0
 //in IA SIMD no greater-than-or-equal comparison for integers,
 // there is greater-than available only, so we need the following tricks
 
-uint8x8_t vcge_s8(int8x8_t a,  int8x8_t b); // VCGE.S8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vcge_s8(int8x8_t a,  int8x8_t b); // VCGE.S8 d0, d0, d0
 _NEON2SSE_INLINE int8x8_t vcge_s8(int8x8_t a,  int8x8_t b)
 {
     int8x8_t res64;
@@ -5117,7 +5119,7 @@ _NEON2SSE_INLINE int8x8_t vcge_s8(int8x8_t a,  int8x8_t b)
 }
 
 
-uint16x4_t vcge_s16(int16x4_t a,  int16x4_t b); // VCGE.S16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcge_s16(int16x4_t a,  int16x4_t b); // VCGE.S16 d0, d0, d0
 _NEON2SSE_INLINE int16x4_t vcge_s16(int16x4_t a,  int16x4_t b)
 {
     int16x4_t res64;
@@ -5125,7 +5127,7 @@ _NEON2SSE_INLINE int16x4_t vcge_s16(int16x4_t a,  int16x4_t b)
 }
 
 
-uint32x2_t vcge_s32(int32x2_t a,  int32x2_t b); // VCGE.S32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcge_s32(int32x2_t a,  int32x2_t b); // VCGE.S32 d0, d0, d0
 _NEON2SSE_INLINE int32x2_t vcge_s32(int32x2_t a,  int32x2_t b)
 {
     int32x2_t res64;
@@ -5133,7 +5135,7 @@ _NEON2SSE_INLINE int32x2_t vcge_s32(int32x2_t a,  int32x2_t b)
 }
 
 
-uint32x2_t vcge_f32(float32x2_t a, float32x2_t b); // VCGE.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcge_f32(float32x2_t a, float32x2_t b); // VCGE.F32 d0, d0, d0
 _NEON2SSE_INLINE uint32x2_t vcge_f32(float32x2_t a, float32x2_t b)
 {
     uint32x2_t res64;
@@ -5142,7 +5144,7 @@ _NEON2SSE_INLINE uint32x2_t vcge_f32(float32x2_t a, float32x2_t b)
     return64f(res);
 }
 
-uint8x8_t vcge_u8(uint8x8_t a,  uint8x8_t b); // VCGE.U8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vcge_u8(uint8x8_t a,  uint8x8_t b); // VCGE.U8 d0, d0, d0
 _NEON2SSE_INLINE uint8x8_t vcge_u8(uint8x8_t a,  uint8x8_t b)
 {
     uint8x8_t res64;
@@ -5150,7 +5152,7 @@ _NEON2SSE_INLINE uint8x8_t vcge_u8(uint8x8_t a,  uint8x8_t b)
 }
 
 
-uint16x4_t vcge_u16(uint16x4_t a,  uint16x4_t b); // VCGE.s16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcge_u16(uint16x4_t a,  uint16x4_t b); // VCGE.s16 d0, d0, d0
 _NEON2SSE_INLINE uint16x4_t vcge_u16(uint16x4_t a,  uint16x4_t b)
 {
     uint16x4_t res64;
@@ -5158,7 +5160,7 @@ _NEON2SSE_INLINE uint16x4_t vcge_u16(uint16x4_t a,  uint16x4_t b)
 }
 
 
-uint32x2_t vcge_u32(uint32x2_t a,  uint32x2_t b); // VCGE.U32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcge_u32(uint32x2_t a,  uint32x2_t b); // VCGE.U32 d0, d0, d0
 _NEON2SSE_INLINE uint32x2_t vcge_u32(uint32x2_t a,  uint32x2_t b)
 {
     //serial solution looks faster
@@ -5168,7 +5170,7 @@ _NEON2SSE_INLINE uint32x2_t vcge_u32(uint32x2_t a,  uint32x2_t b)
 
 
 
-uint8x16_t vcgeq_s8(int8x16_t a, int8x16_t b); // VCGE.S8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vcgeq_s8(int8x16_t a, int8x16_t b); // VCGE.S8 q0, q0, q0
 _NEON2SSE_INLINE uint8x16_t vcgeq_s8(int8x16_t a, int8x16_t b) // VCGE.S8 q0, q0, q0
 {
     __m128i m1, m2;
@@ -5177,7 +5179,7 @@ _NEON2SSE_INLINE uint8x16_t vcgeq_s8(int8x16_t a, int8x16_t b) // VCGE.S8 q0, q0
     return _mm_or_si128  ( m1, m2);
 }
 
-uint16x8_t vcgeq_s16(int16x8_t a, int16x8_t b); // VCGE.S16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcgeq_s16(int16x8_t a, int16x8_t b); // VCGE.S16 q0, q0, q0
 _NEON2SSE_INLINE uint16x8_t vcgeq_s16(int16x8_t a, int16x8_t b) // VCGE.S16 q0, q0, q0
 {
     __m128i m1, m2;
@@ -5186,7 +5188,7 @@ _NEON2SSE_INLINE uint16x8_t vcgeq_s16(int16x8_t a, int16x8_t b) // VCGE.S16 q0,
     return _mm_or_si128   ( m1,m2);
 }
 
-uint32x4_t vcgeq_s32(int32x4_t a, int32x4_t b); // VCGE.S32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcgeq_s32(int32x4_t a, int32x4_t b); // VCGE.S32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vcgeq_s32(int32x4_t a, int32x4_t b) // VCGE.S32 q0, q0, q0
 {
     __m128i m1, m2;
@@ -5195,7 +5197,7 @@ _NEON2SSE_INLINE uint32x4_t vcgeq_s32(int32x4_t a, int32x4_t b) // VCGE.S32 q0,
     return _mm_or_si128   (m1, m2);
 }
 
-uint32x4_t vcgeq_f32(float32x4_t a, float32x4_t b); // VCGE.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcgeq_f32(float32x4_t a, float32x4_t b); // VCGE.F32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vcgeq_f32(float32x4_t a, float32x4_t b)
 {
     __m128 res;
@@ -5203,58 +5205,58 @@ _NEON2SSE_INLINE uint32x4_t vcgeq_f32(float32x4_t a, float32x4_t b)
     return *(__m128i*)&res;
 }
 
-uint8x16_t vcgeq_u8(uint8x16_t a, uint8x16_t b); // VCGE.U8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vcgeq_u8(uint8x16_t a, uint8x16_t b); // VCGE.U8 q0, q0, q0
 _NEON2SSE_INLINE uint8x16_t vcgeq_u8(uint8x16_t a, uint8x16_t b) // VCGE.U8 q0, q0, q0
 {
     //no unsigned chars comparison, only signed available,so need the trick
-        __m128i cmp;
-        cmp = _mm_max_epu8(a, b);
-        return _mm_cmpeq_epi8(cmp, a); //a>=b
+    __m128i cmp;
+    cmp = _mm_max_epu8(a, b);
+    return _mm_cmpeq_epi8(cmp, a); //a>=b
 }
 
-uint16x8_t vcgeq_u16(uint16x8_t a, uint16x8_t b); // VCGE.s16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcgeq_u16(uint16x8_t a, uint16x8_t b); // VCGE.s16 q0, q0, q0
 _NEON2SSE_INLINE uint16x8_t vcgeq_u16(uint16x8_t a, uint16x8_t b) // VCGE.s16 q0, q0, q0
 {
     //no unsigned shorts comparison, only signed available,so need the trick
-    #ifdef USE_SSE4
-        __m128i cmp;
-        cmp = _mm_max_epu16(a, b);
-        return _mm_cmpeq_epi16(cmp, a); //a>=b
-    #else
-	__m128i as, mask;
-	__m128i zero = _mm_setzero_si128();
-	__m128i cffff = _mm_set1_epi16(0xffff);
-	as = _mm_subs_epu16(b,a);
-	mask = _mm_cmpgt_epi16(as, zero);
-	return _mm_xor_si128 ( mask, cffff);
-    #endif
+#ifdef USE_SSE4
+    __m128i cmp;
+    cmp = _mm_max_epu16(a, b);
+    return _mm_cmpeq_epi16(cmp, a); //a>=b
+#else
+    __m128i as, mask;
+    __m128i zero = _mm_setzero_si128();
+    __m128i cffff = _mm_set1_epi16(0xffff);
+    as = _mm_subs_epu16(b,a);
+    mask = _mm_cmpgt_epi16(as, zero);
+    return _mm_xor_si128 ( mask, cffff);
+#endif
 }
 
-uint32x4_t vcgeq_u32(uint32x4_t a, uint32x4_t b); // VCGE.U32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcgeq_u32(uint32x4_t a, uint32x4_t b); // VCGE.U32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vcgeq_u32(uint32x4_t a, uint32x4_t b) // VCGE.U32 q0, q0, q0
 {
     //no unsigned ints comparison, only signed available,so need the trick
-    #ifdef USE_SSE4
-        __m128i cmp;
-        cmp = _mm_max_epu32(a, b);
-        return _mm_cmpeq_epi32(cmp, a); //a>=b
-    #else
-        //serial solution may be faster
-        __m128i c80000000, as, bs, m1, m2;
-        c80000000 = _mm_set1_epi32 (0x80000000);
-        as = _mm_sub_epi32(a,c80000000);
-        bs = _mm_sub_epi32(b,c80000000);
-        m1 = _mm_cmpgt_epi32 (as, bs);
-        m2 = _mm_cmpeq_epi32 (as, bs);
-        return _mm_or_si128 ( m1,  m2);
-    #endif
+#ifdef USE_SSE4
+    __m128i cmp;
+    cmp = _mm_max_epu32(a, b);
+    return _mm_cmpeq_epi32(cmp, a); //a>=b
+#else
+    //serial solution may be faster
+    __m128i c80000000, as, bs, m1, m2;
+    c80000000 = _mm_set1_epi32 (0x80000000);
+    as = _mm_sub_epi32(a,c80000000);
+    bs = _mm_sub_epi32(b,c80000000);
+    m1 = _mm_cmpgt_epi32 (as, bs);
+    m2 = _mm_cmpeq_epi32 (as, bs);
+    return _mm_or_si128 ( m1,  m2);
+#endif
 }
 
 //**********************Vector compare less-than or equal******************************
 //***************************************************************************************
 //in IA SIMD no less-than-or-equal comparison for integers present, so we need the tricks
 
-uint8x8_t vcle_s8(int8x8_t a,  int8x8_t b); // VCGE.S8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vcle_s8(int8x8_t a,  int8x8_t b); // VCGE.S8 d0, d0, d0
 _NEON2SSE_INLINE int8x8_t vcle_s8(int8x8_t a,  int8x8_t b)
 {
     int8x8_t res64;
@@ -5262,7 +5264,7 @@ _NEON2SSE_INLINE int8x8_t vcle_s8(int8x8_t a,  int8x8_t b)
 }
 
 
-uint16x4_t vcle_s16(int16x4_t a,  int16x4_t b); // VCGE.S16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcle_s16(int16x4_t a,  int16x4_t b); // VCGE.S16 d0, d0, d0
 _NEON2SSE_INLINE int16x4_t vcle_s16(int16x4_t a,  int16x4_t b)
 {
     int16x4_t res64;
@@ -5270,7 +5272,7 @@ _NEON2SSE_INLINE int16x4_t vcle_s16(int16x4_t a,  int16x4_t b)
 }
 
 
-uint32x2_t vcle_s32(int32x2_t a,  int32x2_t b); // VCGE.S32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcle_s32(int32x2_t a,  int32x2_t b); // VCGE.S32 d0, d0, d0
 _NEON2SSE_INLINE int32x2_t vcle_s32(int32x2_t a,  int32x2_t b)
 {
     int32x2_t res64;
@@ -5278,7 +5280,7 @@ _NEON2SSE_INLINE int32x2_t vcle_s32(int32x2_t a,  int32x2_t b)
 }
 
 
-uint32x2_t vcle_f32(float32x2_t a, float32x2_t b); // VCGE.F32 d0, d0, d0?
+_NEON2SSESTORAGE uint32x2_t vcle_f32(float32x2_t a, float32x2_t b); // VCGE.F32 d0, d0, d0?
 _NEON2SSE_INLINE uint32x2_t vcle_f32(float32x2_t a, float32x2_t b)
 {
     uint32x2_t res64;
@@ -5287,18 +5289,18 @@ _NEON2SSE_INLINE uint32x2_t vcle_f32(float32x2_t a, float32x2_t b)
     return64f(res);
 }
 
-uint8x8_t vcle_u8(uint8x8_t a,  uint8x8_t b); // VCGE.U8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vcle_u8(uint8x8_t a,  uint8x8_t b); // VCGE.U8 d0, d0, d0
 #define vcle_u8(a,b) vcge_u8(b,a)
 
 
-uint16x4_t vcle_u16(uint16x4_t a,  uint16x4_t b); // VCGE.s16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcle_u16(uint16x4_t a,  uint16x4_t b); // VCGE.s16 d0, d0, d0
 #define vcle_u16(a,b) vcge_u16(b,a)
 
 
-uint32x2_t vcle_u32(uint32x2_t a,  uint32x2_t b); // VCGE.U32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcle_u32(uint32x2_t a,  uint32x2_t b); // VCGE.U32 d0, d0, d0
 #define vcle_u32(a,b) vcge_u32(b,a)
 
-uint8x16_t vcleq_s8(int8x16_t a, int8x16_t b); // VCGE.S8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vcleq_s8(int8x16_t a, int8x16_t b); // VCGE.S8 q0, q0, q0
 _NEON2SSE_INLINE uint8x16_t vcleq_s8(int8x16_t a, int8x16_t b) // VCGE.S8 q0, q0, q0
 {
     __m128i c1, res;
@@ -5307,7 +5309,7 @@ _NEON2SSE_INLINE uint8x16_t vcleq_s8(int8x16_t a, int8x16_t b) // VCGE.S8 q0, q0
     return _mm_andnot_si128 (res, c1); //inverse the cmpgt result, get less-than-or-equal
 }
 
-uint16x8_t vcleq_s16(int16x8_t a, int16x8_t b); // VCGE.S16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcleq_s16(int16x8_t a, int16x8_t b); // VCGE.S16 q0, q0, q0
 _NEON2SSE_INLINE uint16x8_t vcleq_s16(int16x8_t a, int16x8_t b) // VCGE.S16 q0, q0, q0
 {
     __m128i c1, res;
@@ -5316,7 +5318,7 @@ _NEON2SSE_INLINE uint16x8_t vcleq_s16(int16x8_t a, int16x8_t b) // VCGE.S16 q0,
     return _mm_andnot_si128 (res, c1);
 }
 
-uint32x4_t vcleq_s32(int32x4_t a, int32x4_t b); // VCGE.S32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcleq_s32(int32x4_t a, int32x4_t b); // VCGE.S32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vcleq_s32(int32x4_t a, int32x4_t b) // VCGE.S32 q0, q0, q0
 {
     __m128i c1, res;
@@ -5325,7 +5327,7 @@ _NEON2SSE_INLINE uint32x4_t vcleq_s32(int32x4_t a, int32x4_t b) // VCGE.S32 q0,
     return _mm_andnot_si128 (res, c1);
 }
 
-uint32x4_t vcleq_f32(float32x4_t a, float32x4_t b); // VCGE.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcleq_f32(float32x4_t a, float32x4_t b); // VCGE.F32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vcleq_f32(float32x4_t a, float32x4_t b)
 {
     __m128 res;
@@ -5333,7 +5335,7 @@ _NEON2SSE_INLINE uint32x4_t vcleq_f32(float32x4_t a, float32x4_t b)
     return *(__m128i*)&res;
 }
 
-uint8x16_t vcleq_u8(uint8x16_t a, uint8x16_t b); // VCGE.U8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vcleq_u8(uint8x16_t a, uint8x16_t b); // VCGE.U8 q0, q0, q0
 #ifdef USE_SSE4
     _NEON2SSE_INLINE uint8x16_t vcleq_u8(uint8x16_t a, uint8x16_t b) // VCGE.U8 q0, q0, q0
     {
@@ -5343,11 +5345,11 @@ uint8x16_t vcleq_u8(uint8x16_t a, uint8x16_t b); // VCGE.U8 q0, q0, q0
         return _mm_cmpeq_epi8(cmp, a); //a<=b
     }
 #else
-    #define vcleq_u8(a,b) vcgeq_u8(b,a)
+#   define vcleq_u8(a,b) vcgeq_u8(b,a)
 #endif
 
 
-uint16x8_t vcleq_u16(uint16x8_t a, uint16x8_t b); // VCGE.s16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcleq_u16(uint16x8_t a, uint16x8_t b); // VCGE.s16 q0, q0, q0
 #ifdef USE_SSE4
     _NEON2SSE_INLINE uint16x8_t vcleq_u16(uint16x8_t a, uint16x8_t b) // VCGE.s16 q0, q0, q0
     {
@@ -5357,11 +5359,11 @@ uint16x8_t vcleq_u16(uint16x8_t a, uint16x8_t b); // VCGE.s16 q0, q0, q0
         return _mm_cmpeq_epi16(cmp, a); //a<=b
     }
 #else
-    #define vcleq_u16(a,b) vcgeq_u16(b,a)
+#   define vcleq_u16(a,b) vcgeq_u16(b,a)
 #endif
 
 
-uint32x4_t vcleq_u32(uint32x4_t a, uint32x4_t b); // VCGE.U32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcleq_u32(uint32x4_t a, uint32x4_t b); // VCGE.U32 q0, q0, q0
 #ifdef USE_SSE4
     _NEON2SSE_INLINE uint32x4_t vcleq_u32(uint32x4_t a, uint32x4_t b) // VCGE.U32 q0, q0, q0
     {
@@ -5372,13 +5374,13 @@ uint32x4_t vcleq_u32(uint32x4_t a, uint32x4_t b); // VCGE.U32 q0, q0, q0
     }
 #else
 //solution may be not optimal compared with the serial one
-    #define vcleq_u32(a,b) vcgeq_u32(b,a)
+#   define vcleq_u32(a,b) vcgeq_u32(b,a)
 #endif
 
 
 //****** Vector compare greater-than ******************************************
 //**************************************************************************
-uint8x8_t vcgt_s8(int8x8_t a, int8x8_t b); // VCGT.S8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vcgt_s8(int8x8_t a, int8x8_t b); // VCGT.S8 d0, d0, d0
 _NEON2SSE_INLINE int8x8_t vcgt_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
@@ -5386,7 +5388,7 @@ _NEON2SSE_INLINE int8x8_t vcgt_s8(int8x8_t a, int8x8_t b)
 }
 
 
-uint16x4_t vcgt_s16(int16x4_t a, int16x4_t b); // VCGT.S16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcgt_s16(int16x4_t a, int16x4_t b); // VCGT.S16 d0, d0, d0
 _NEON2SSE_INLINE int16x4_t vcgt_s16(int16x4_t a, int16x4_t b)
 {
     int16x4_t res64;
@@ -5394,7 +5396,7 @@ _NEON2SSE_INLINE int16x4_t vcgt_s16(int16x4_t a, int16x4_t b)
 }
 
 
-uint32x2_t vcgt_s32(int32x2_t a, int32x2_t b); // VCGT.S32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcgt_s32(int32x2_t a, int32x2_t b); // VCGT.S32 d0, d0, d0
 _NEON2SSE_INLINE int32x2_t vcgt_s32(int32x2_t a, int32x2_t b)
 {
     int32x2_t res64;
@@ -5402,7 +5404,7 @@ _NEON2SSE_INLINE int32x2_t vcgt_s32(int32x2_t a, int32x2_t b)
 }
 
 
-uint32x2_t vcgt_f32(float32x2_t a, float32x2_t b); // VCGT.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcgt_f32(float32x2_t a, float32x2_t b); // VCGT.F32 d0, d0, d0
 _NEON2SSE_INLINE uint32x2_t vcgt_f32(float32x2_t a, float32x2_t b)
 {
     uint32x2_t res64;
@@ -5411,7 +5413,7 @@ _NEON2SSE_INLINE uint32x2_t vcgt_f32(float32x2_t a, float32x2_t b)
     return64f(res);
 }
 
-uint8x8_t vcgt_u8(uint8x8_t a,  uint8x8_t b); // VCGT.U8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vcgt_u8(uint8x8_t a,  uint8x8_t b); // VCGT.U8 d0, d0, d0
 _NEON2SSE_INLINE uint8x8_t vcgt_u8(uint8x8_t a,  uint8x8_t b)
 {
     uint8x8_t res64;
@@ -5419,7 +5421,7 @@ _NEON2SSE_INLINE uint8x8_t vcgt_u8(uint8x8_t a,  uint8x8_t b)
 }
 
 
-uint16x4_t vcgt_u16(uint16x4_t a,  uint16x4_t b); // VCGT.s16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vcgt_u16(uint16x4_t a,  uint16x4_t b); // VCGT.s16 d0, d0, d0
 _NEON2SSE_INLINE uint16x4_t vcgt_u16(uint16x4_t a,  uint16x4_t b)
 {
     uint16x4_t res64;
@@ -5427,7 +5429,7 @@ _NEON2SSE_INLINE uint16x4_t vcgt_u16(uint16x4_t a,  uint16x4_t b)
 }
 
 
-uint32x2_t vcgt_u32(uint32x2_t a,  uint32x2_t b); // VCGT.U32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcgt_u32(uint32x2_t a,  uint32x2_t b); // VCGT.U32 d0, d0, d0
 _NEON2SSE_INLINE uint32x2_t vcgt_u32(uint32x2_t a,  uint32x2_t b)
 {
     uint32x2_t res64;
@@ -5435,16 +5437,16 @@ _NEON2SSE_INLINE uint32x2_t vcgt_u32(uint32x2_t a,  uint32x2_t b)
 }
 
 
-uint8x16_t   vcgtq_s8(int8x16_t a, int8x16_t b); // VCGT.S8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t   vcgtq_s8(int8x16_t a, int8x16_t b); // VCGT.S8 q0, q0, q0
 #define vcgtq_s8 _mm_cmpgt_epi8
 
-uint16x8_t   vcgtq_s16(int16x8_t a, int16x8_t b); // VCGT.S16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t   vcgtq_s16(int16x8_t a, int16x8_t b); // VCGT.S16 q0, q0, q0
 #define vcgtq_s16 _mm_cmpgt_epi16
 
-uint32x4_t   vcgtq_s32(int32x4_t a, int32x4_t b); // VCGT.S32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t   vcgtq_s32(int32x4_t a, int32x4_t b); // VCGT.S32 q0, q0, q0
 #define vcgtq_s32 _mm_cmpgt_epi32
 
-uint32x4_t vcgtq_f32(float32x4_t a, float32x4_t b); // VCGT.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcgtq_f32(float32x4_t a, float32x4_t b); // VCGT.F32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vcgtq_f32(float32x4_t a, float32x4_t b)
 {
     __m128 res;
@@ -5452,27 +5454,27 @@ _NEON2SSE_INLINE uint32x4_t vcgtq_f32(float32x4_t a, float32x4_t b)
     return *(__m128i*)&res;
 }
 
-uint8x16_t vcgtq_u8(uint8x16_t a, uint8x16_t b); // VCGT.U8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vcgtq_u8(uint8x16_t a, uint8x16_t b); // VCGT.U8 q0, q0, q0
 _NEON2SSE_INLINE uint8x16_t vcgtq_u8(uint8x16_t a, uint8x16_t b) // VCGT.U8 q0, q0, q0
 {
     //no unsigned chars comparison, only signed available,so need the trick
-	__m128i as, mask;
-	__m128i zero = _mm_setzero_si128();
-	as = _mm_subs_epu8(a, b);
-	return _mm_cmpgt_epi8(as, zero);
+    __m128i as;
+    __m128i zero = _mm_setzero_si128();
+    as = _mm_subs_epu8(a, b);
+    return _mm_cmpgt_epi8(as, zero);
 }
 
-uint16x8_t vcgtq_u16(uint16x8_t a, uint16x8_t b); // VCGT.s16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcgtq_u16(uint16x8_t a, uint16x8_t b); // VCGT.s16 q0, q0, q0
 _NEON2SSE_INLINE uint16x8_t vcgtq_u16(uint16x8_t a, uint16x8_t b) // VCGT.s16 q0, q0, q0
 {
     //no unsigned short comparison, only signed available,so need the trick
-	__m128i as, mask;
-	__m128i zero = _mm_setzero_si128();
-	as = _mm_subs_epu16(a, b);
-	return _mm_cmpgt_epi16(as, zero);
+    __m128i as;
+    __m128i zero = _mm_setzero_si128();
+    as = _mm_subs_epu16(a, b);
+    return _mm_cmpgt_epi16(as, zero);
 }
 
-uint32x4_t vcgtq_u32(uint32x4_t a, uint32x4_t b); // VCGT.U32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcgtq_u32(uint32x4_t a, uint32x4_t b); // VCGT.U32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vcgtq_u32(uint32x4_t a, uint32x4_t b) // VCGT.U32 q0, q0, q0
 {
     //no unsigned int comparison, only signed available,so need the trick
@@ -5485,54 +5487,54 @@ _NEON2SSE_INLINE uint32x4_t vcgtq_u32(uint32x4_t a, uint32x4_t b) // VCGT.U32 q0
 
 //********************* Vector compare less-than **************************
 //*************************************************************************
-uint8x8_t   vclt_s8(int8x8_t a, int8x8_t b); // VCGT.S8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t   vclt_s8(int8x8_t a, int8x8_t b); // VCGT.S8 d0, d0, d0
 #define vclt_s8(a,b) vcgt_s8(b,a) //swap the arguments!!
 
 
-uint16x4_t   vclt_s16(int16x4_t a, int16x4_t b); // VCGT.S16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t   vclt_s16(int16x4_t a, int16x4_t b); // VCGT.S16 d0, d0, d0
 #define vclt_s16(a,b) vcgt_s16(b,a) //swap the arguments!!
 
 
-uint32x2_t   vclt_s32(int32x2_t a, int32x2_t b); // VCGT.S32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t   vclt_s32(int32x2_t a, int32x2_t b); // VCGT.S32 d0, d0, d0
 #define vclt_s32(a,b)  vcgt_s32(b,a) //swap the arguments!!
 
 
-uint32x2_t vclt_f32(float32x2_t a, float32x2_t b); // VCGT.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vclt_f32(float32x2_t a, float32x2_t b); // VCGT.F32 d0, d0, d0
 #define vclt_f32(a,b) vcgt_f32(b, a) //swap the arguments!!
 
-uint8x8_t vclt_u8(uint8x8_t a, uint8x8_t b); // VCGT.U8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vclt_u8(uint8x8_t a, uint8x8_t b); // VCGT.U8 d0, d0, d0
 #define vclt_u8(a,b) vcgt_u8(b,a) //swap the arguments!!
 
-uint16x4_t vclt_u16(uint16x4_t a, uint16x4_t b); // VCGT.s16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vclt_u16(uint16x4_t a, uint16x4_t b); // VCGT.s16 d0, d0, d0
 #define vclt_u16(a,b) vcgt_u16(b,a) //swap the arguments!!
 
-uint32x2_t vclt_u32(uint32x2_t a, uint32x2_t b); // VCGT.U32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vclt_u32(uint32x2_t a, uint32x2_t b); // VCGT.U32 d0, d0, d0
 #define vclt_u32(a,b) vcgt_u32(b,a) //swap the arguments!!
 
-uint8x16_t   vcltq_s8(int8x16_t a, int8x16_t b); // VCGT.S8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t   vcltq_s8(int8x16_t a, int8x16_t b); // VCGT.S8 q0, q0, q0
 #define vcltq_s8(a,b) vcgtq_s8(b, a) //swap the arguments!!
 
-uint16x8_t   vcltq_s16(int16x8_t a, int16x8_t b); // VCGT.S16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t   vcltq_s16(int16x8_t a, int16x8_t b); // VCGT.S16 q0, q0, q0
 #define vcltq_s16(a,b) vcgtq_s16(b, a) //swap the arguments!!
 
-uint32x4_t   vcltq_s32(int32x4_t a, int32x4_t b); // VCGT.S32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t   vcltq_s32(int32x4_t a, int32x4_t b); // VCGT.S32 q0, q0, q0
 #define vcltq_s32(a,b) vcgtq_s32(b, a) //swap the arguments!!
 
-uint32x4_t vcltq_f32(float32x4_t a, float32x4_t b); // VCGT.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcltq_f32(float32x4_t a, float32x4_t b); // VCGT.F32 q0, q0, q0
 #define vcltq_f32(a,b) vcgtq_f32(b, a) //swap the arguments!!
 
-uint8x16_t vcltq_u8(uint8x16_t a, uint8x16_t b); // VCGT.U8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vcltq_u8(uint8x16_t a, uint8x16_t b); // VCGT.U8 q0, q0, q0
 #define vcltq_u8(a,b) vcgtq_u8(b, a) //swap the arguments!!
 
-uint16x8_t vcltq_u16(uint16x8_t a, uint16x8_t b); // VCGT.s16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vcltq_u16(uint16x8_t a, uint16x8_t b); // VCGT.s16 q0, q0, q0
 #define vcltq_u16(a,b) vcgtq_u16(b, a) //swap the arguments!!
 
-uint32x4_t vcltq_u32(uint32x4_t a, uint32x4_t b); // VCGT.U32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcltq_u32(uint32x4_t a, uint32x4_t b); // VCGT.U32 q0, q0, q0
 #define vcltq_u32(a,b) vcgtq_u32(b, a) //swap the arguments!!
 
 //*****************Vector compare absolute greater-than or equal ************
 //***************************************************************************
-uint32x2_t vcage_f32(float32x2_t a, float32x2_t b); // VACGE.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcage_f32(float32x2_t a, float32x2_t b); // VACGE.F32 d0, d0, d0
 _NEON2SSE_INLINE uint32x2_t vcage_f32(float32x2_t a, float32x2_t b)
 {
     uint32x2_t res64;
@@ -5545,7 +5547,7 @@ _NEON2SSE_INLINE uint32x2_t vcage_f32(float32x2_t a, float32x2_t b)
     return64f(a0);
 }
 
-uint32x4_t vcageq_f32(float32x4_t a, float32x4_t b); // VACGE.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcageq_f32(float32x4_t a, float32x4_t b); // VACGE.F32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vcageq_f32(float32x4_t a, float32x4_t b) // VACGE.F32 q0, q0, q0
 {
     __m128i c7fffffff;
@@ -5559,7 +5561,7 @@ _NEON2SSE_INLINE uint32x4_t vcageq_f32(float32x4_t a, float32x4_t b) // VACGE.F3
 
 //********Vector compare absolute less-than or equal ******************
 //********************************************************************
-uint32x2_t vcale_f32(float32x2_t a, float32x2_t b); // VACGE.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcale_f32(float32x2_t a, float32x2_t b); // VACGE.F32 d0, d0, d0
 _NEON2SSE_INLINE uint32x2_t vcale_f32(float32x2_t a, float32x2_t b)
 {
     uint32x2_t res64;
@@ -5572,7 +5574,7 @@ _NEON2SSE_INLINE uint32x2_t vcale_f32(float32x2_t a, float32x2_t b)
     return64f(a0);
 }
 
-uint32x4_t vcaleq_f32(float32x4_t a, float32x4_t b); // VACGE.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcaleq_f32(float32x4_t a, float32x4_t b); // VACGE.F32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vcaleq_f32(float32x4_t a, float32x4_t b) // VACGE.F32 q0, q0, q0
 {
     __m128i c7fffffff;
@@ -5586,7 +5588,7 @@ _NEON2SSE_INLINE uint32x4_t vcaleq_f32(float32x4_t a, float32x4_t b) // VACGE.F3
 
 //********  Vector compare absolute greater-than    ******************
 //******************************************************************
-uint32x2_t vcagt_f32(float32x2_t a, float32x2_t b); // VACGT.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcagt_f32(float32x2_t a, float32x2_t b); // VACGT.F32 d0, d0, d0
 _NEON2SSE_INLINE uint32x2_t vcagt_f32(float32x2_t a, float32x2_t b)
 {
     uint32x2_t res64;
@@ -5599,7 +5601,7 @@ _NEON2SSE_INLINE uint32x2_t vcagt_f32(float32x2_t a, float32x2_t b)
     return64f(a0);
 }
 
-uint32x4_t vcagtq_f32(float32x4_t a, float32x4_t b); // VACGT.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcagtq_f32(float32x4_t a, float32x4_t b); // VACGT.F32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vcagtq_f32(float32x4_t a, float32x4_t b) // VACGT.F32 q0, q0, q0
 {
     __m128i c7fffffff;
@@ -5613,7 +5615,7 @@ _NEON2SSE_INLINE uint32x4_t vcagtq_f32(float32x4_t a, float32x4_t b) // VACGT.F3
 
 //***************Vector compare absolute less-than  ***********************
 //*************************************************************************
-uint32x2_t vcalt_f32(float32x2_t a, float32x2_t b); // VACGT.F32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vcalt_f32(float32x2_t a, float32x2_t b); // VACGT.F32 d0, d0, d0
 _NEON2SSE_INLINE uint32x2_t vcalt_f32(float32x2_t a, float32x2_t b)
 {
     uint32x2_t res64;
@@ -5626,7 +5628,7 @@ _NEON2SSE_INLINE uint32x2_t vcalt_f32(float32x2_t a, float32x2_t b)
     return64f(a0);
 }
 
-uint32x4_t vcaltq_f32(float32x4_t a, float32x4_t b); // VACGT.F32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vcaltq_f32(float32x4_t a, float32x4_t b); // VACGT.F32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vcaltq_f32(float32x4_t a, float32x4_t b) // VACGT.F32 q0, q0, q0
 {
     __m128i c7fffffff;
@@ -5645,7 +5647,7 @@ with the corresponding element of a second vector. If the result is not zero, th
 corresponding element in the destination vector is set to all ones. Otherwise, it is set to
 all zeros. */
 
-uint8x8_t vtst_s8(int8x8_t a,  int8x8_t b); // VTST.8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vtst_s8(int8x8_t a,  int8x8_t b); // VTST.8 d0, d0, d0
 _NEON2SSE_INLINE uint8x8_t vtst_s8(int8x8_t a,  int8x8_t b)
 {
     int8x8_t res64;
@@ -5653,7 +5655,7 @@ _NEON2SSE_INLINE uint8x8_t vtst_s8(int8x8_t a,  int8x8_t b)
 }
 
 
-uint16x4_t vtst_s16(int16x4_t a,  int16x4_t b); // VTST.16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vtst_s16(int16x4_t a,  int16x4_t b); // VTST.16 d0, d0, d0
 _NEON2SSE_INLINE uint16x4_t vtst_s16(int16x4_t a,  int16x4_t b)
 {
     int16x4_t res64;
@@ -5661,7 +5663,7 @@ _NEON2SSE_INLINE uint16x4_t vtst_s16(int16x4_t a,  int16x4_t b)
 }
 
 
-uint32x2_t vtst_s32(int32x2_t a,  int32x2_t b); // VTST.32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vtst_s32(int32x2_t a,  int32x2_t b); // VTST.32 d0, d0, d0
 _NEON2SSE_INLINE uint32x2_t vtst_s32(int32x2_t a,  int32x2_t b)
 {
     int32x2_t res64;
@@ -5669,20 +5671,20 @@ _NEON2SSE_INLINE uint32x2_t vtst_s32(int32x2_t a,  int32x2_t b)
 }
 
 
-uint8x8_t vtst_u8(uint8x8_t a,  uint8x8_t b); // VTST.8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vtst_u8(uint8x8_t a,  uint8x8_t b); // VTST.8 d0, d0, d0
 #define vtst_u8 vtst_s8
 
-uint16x4_t vtst_u16(uint16x4_t a,  uint16x4_t b); // VTST.16 d0, d0, d0
+_NEON2SSESTORAGE uint16x4_t vtst_u16(uint16x4_t a,  uint16x4_t b); // VTST.16 d0, d0, d0
 #define vtst_u16 vtst_s16
 
-uint32x2_t vtst_u32(uint32x2_t a,  uint32x2_t b); // VTST.32 d0, d0, d0
+_NEON2SSESTORAGE uint32x2_t vtst_u32(uint32x2_t a,  uint32x2_t b); // VTST.32 d0, d0, d0
 #define vtst_u32 vtst_s32
 
 
-uint8x8_t vtst_p8(poly8x8_t a, poly8x8_t b); // VTST.8 d0, d0, d0
+_NEON2SSESTORAGE uint8x8_t vtst_p8(poly8x8_t a, poly8x8_t b); // VTST.8 d0, d0, d0
 #define vtst_p8 vtst_u8
 
-uint8x16_t vtstq_s8(int8x16_t a, int8x16_t b); // VTST.8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vtstq_s8(int8x16_t a, int8x16_t b); // VTST.8 q0, q0, q0
 _NEON2SSE_INLINE uint8x16_t vtstq_s8(int8x16_t a, int8x16_t b) // VTST.8 q0, q0, q0
 {
     __m128i zero, one, res;
@@ -5693,7 +5695,7 @@ _NEON2SSE_INLINE uint8x16_t vtstq_s8(int8x16_t a, int8x16_t b) // VTST.8 q0, q0,
     return _mm_xor_si128(res, one); //invert result
 }
 
-uint16x8_t vtstq_s16(int16x8_t a, int16x8_t b); // VTST.16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vtstq_s16(int16x8_t a, int16x8_t b); // VTST.16 q0, q0, q0
 _NEON2SSE_INLINE uint16x8_t vtstq_s16(int16x8_t a, int16x8_t b) // VTST.16 q0, q0, q0
 {
     __m128i zero, one, res;
@@ -5704,7 +5706,7 @@ _NEON2SSE_INLINE uint16x8_t vtstq_s16(int16x8_t a, int16x8_t b) // VTST.16 q0, q
     return _mm_xor_si128(res, one); //invert result
 }
 
-uint32x4_t vtstq_s32(int32x4_t a, int32x4_t b); // VTST.32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vtstq_s32(int32x4_t a, int32x4_t b); // VTST.32 q0, q0, q0
 _NEON2SSE_INLINE uint32x4_t vtstq_s32(int32x4_t a, int32x4_t b) // VTST.32 q0, q0, q0
 {
     __m128i zero, one, res;
@@ -5715,36 +5717,36 @@ _NEON2SSE_INLINE uint32x4_t vtstq_s32(int32x4_t a, int32x4_t b) // VTST.32 q0, q
     return _mm_xor_si128(res, one); //invert result
 }
 
-uint8x16_t vtstq_u8(uint8x16_t a, uint8x16_t b); // VTST.8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vtstq_u8(uint8x16_t a, uint8x16_t b); // VTST.8 q0, q0, q0
 #define vtstq_u8 vtstq_s8
 
-uint16x8_t vtstq_u16(uint16x8_t a, uint16x8_t b); // VTST.16 q0, q0, q0
+_NEON2SSESTORAGE uint16x8_t vtstq_u16(uint16x8_t a, uint16x8_t b); // VTST.16 q0, q0, q0
 #define vtstq_u16 vtstq_s16
 
-uint32x4_t vtstq_u32(uint32x4_t a, uint32x4_t b); // VTST.32 q0, q0, q0
+_NEON2SSESTORAGE uint32x4_t vtstq_u32(uint32x4_t a, uint32x4_t b); // VTST.32 q0, q0, q0
 #define vtstq_u32 vtstq_s32
 
-uint8x16_t vtstq_p8(poly8x16_t a, poly8x16_t b); // VTST.8 q0, q0, q0
+_NEON2SSESTORAGE uint8x16_t vtstq_p8(poly8x16_t a, poly8x16_t b); // VTST.8 q0, q0, q0
 #define vtstq_p8 vtstq_u8
 
 //****************** Absolute difference ********************
 //*** Absolute difference between the arguments: Vr[i] = | Va[i] - Vb[i] |*****
 //************************************************************
-int8x8_t vabd_s8(int8x8_t a,  int8x8_t b); // VABD.S8 d0,d0,d0
-_NEON2SSE_INLINE int8x8_t vabd_s8(int8x8_t a,  int8x8_t b) 
-{ 
+_NEON2SSESTORAGE int8x8_t vabd_s8(int8x8_t a,  int8x8_t b); // VABD.S8 d0,d0,d0
+_NEON2SSE_INLINE int8x8_t vabd_s8(int8x8_t a,  int8x8_t b)
+{
     int8x8_t res64;
     return64(vabdq_s8(_pM128i(a), _pM128i(b)));
 }
 
-int16x4_t vabd_s16(int16x4_t a,  int16x4_t b); // VABD.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vabd_s16(int16x4_t a,  int16x4_t b); // VABD.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vabd_s16(int16x4_t a,  int16x4_t b)
 {
     int16x4_t res64;
     return64(vabdq_s16(_pM128i(a), _pM128i(b)));
 }
 
-int32x2_t vabd_s32(int32x2_t a,  int32x2_t b); // VABD.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vabd_s32(int32x2_t a,  int32x2_t b); // VABD.S32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vabd_s32(int32x2_t a,  int32x2_t b)
 {//need to deal with an intermediate overflow
     int32x2_t res;
@@ -5753,28 +5755,28 @@ _NEON2SSE_INLINE int32x2_t vabd_s32(int32x2_t a,  int32x2_t b)
     return res;
 }
 
-uint8x8_t vabd_u8(uint8x8_t a,  uint8x8_t b); // VABD.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vabd_u8(uint8x8_t a,  uint8x8_t b); // VABD.U8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vabd_u8(uint8x8_t a,  uint8x8_t b)
 {
     uint8x8_t res64;
     return64(vabdq_u8(_pM128i(a), _pM128i(b)));
 }
 
-uint16x4_t vabd_u16(uint16x4_t a,  uint16x4_t b); // VABD.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vabd_u16(uint16x4_t a,  uint16x4_t b); // VABD.s16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vabd_u16(uint16x4_t a,  uint16x4_t b)
 {
     uint16x4_t res64;
     return64(vabdq_u16(_pM128i(a), _pM128i(b)));
 }
 
-uint32x2_t vabd_u32(uint32x2_t a,  uint32x2_t b); // VABD.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vabd_u32(uint32x2_t a,  uint32x2_t b); // VABD.U32 d0,d0,d0
 _NEON2SSE_INLINE uint32x2_t vabd_u32(uint32x2_t a,  uint32x2_t b)
 {
     uint32x2_t res64;
     return64(vabdq_u32(_pM128i(a), _pM128i(b)));
 }
 
-float32x2_t vabd_f32(float32x2_t a, float32x2_t b); // VABD.F32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vabd_f32(float32x2_t a, float32x2_t b); // VABD.F32 d0,d0,d0
 _NEON2SSE_INLINE float32x2_t vabd_f32(float32x2_t a, float32x2_t b)
 {
     float32x4_t res;
@@ -5784,43 +5786,43 @@ _NEON2SSE_INLINE float32x2_t vabd_f32(float32x2_t a, float32x2_t b)
     return res64;
 }
 
-int8x16_t vabdq_s8(int8x16_t a, int8x16_t b); // VABD.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vabdq_s8(int8x16_t a, int8x16_t b); // VABD.S8 q0,q0,q0
 _NEON2SSE_INLINE int8x16_t vabdq_s8(int8x16_t a, int8x16_t b) // VABD.S8 q0,q0,q0
 { //need to deal with an intermediate overflow
-   __m128i cmp, difab, difba; 
-   cmp = vcgtq_s8(a,b); 
-   difab = _mm_sub_epi8(a,b); 
-   difba = _mm_sub_epi8(b,a); 
-   difab = _mm_and_si128(cmp, difab); 
-   difba = _mm_andnot_si128(cmp, difba); 
-   return _mm_or_si128(difab, difba); 
+   __m128i cmp, difab, difba;
+   cmp = vcgtq_s8(a,b);
+   difab = _mm_sub_epi8(a,b);
+   difba = _mm_sub_epi8(b,a);
+   difab = _mm_and_si128(cmp, difab);
+   difba = _mm_andnot_si128(cmp, difba);
+   return _mm_or_si128(difab, difba);
 }
 
-int16x8_t vabdq_s16(int16x8_t a, int16x8_t b); // VABD.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vabdq_s16(int16x8_t a, int16x8_t b); // VABD.S16 q0,q0,q0
 _NEON2SSE_INLINE int16x8_t vabdq_s16(int16x8_t a, int16x8_t b) // VABD.S16 q0,q0,q0
 {//need to deal with an intermediate overflow
-    __m128i cmp, difab, difba; 
-    cmp = vcgtq_s16(a,b); 
-    difab = _mm_sub_epi16(a,b); 
-    difba = _mm_sub_epi16 (b,a); 
-    difab = _mm_and_si128(cmp, difab); 
-    difba = _mm_andnot_si128(cmp, difba); 
-    return _mm_or_si128(difab, difba); 
+    __m128i cmp, difab, difba;
+    cmp = vcgtq_s16(a,b);
+    difab = _mm_sub_epi16(a,b);
+    difba = _mm_sub_epi16 (b,a);
+    difab = _mm_and_si128(cmp, difab);
+    difba = _mm_andnot_si128(cmp, difba);
+    return _mm_or_si128(difab, difba);
 }
 
-int32x4_t vabdq_s32(int32x4_t a, int32x4_t b); // VABD.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vabdq_s32(int32x4_t a, int32x4_t b); // VABD.S32 q0,q0,q0
 _NEON2SSE_INLINE int32x4_t vabdq_s32(int32x4_t a, int32x4_t b) // VABD.S32 q0,q0,q0
 {//need to deal with an intermediate overflow
-    __m128i cmp, difab, difba; 
-    cmp = vcgtq_s32(a,b); 
-    difab = _mm_sub_epi32(a,b); 
-    difba = _mm_sub_epi32(b,a); 
-    difab = _mm_and_si128(cmp, difab); 
-    difba = _mm_andnot_si128(cmp, difba); 
-    return _mm_or_si128(difab, difba); 
+    __m128i cmp, difab, difba;
+    cmp = vcgtq_s32(a,b);
+    difab = _mm_sub_epi32(a,b);
+    difba = _mm_sub_epi32(b,a);
+    difab = _mm_and_si128(cmp, difab);
+    difba = _mm_andnot_si128(cmp, difba);
+    return _mm_or_si128(difab, difba);
 }
 
-uint8x16_t vabdq_u8(uint8x16_t a, uint8x16_t b); // VABD.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vabdq_u8(uint8x16_t a, uint8x16_t b); // VABD.U8 q0,q0,q0
 _NEON2SSE_INLINE uint8x16_t vabdq_u8(uint8x16_t a, uint8x16_t b) //no abs for unsigned
 {
     __m128i  difab, difba;
@@ -5829,7 +5831,7 @@ _NEON2SSE_INLINE uint8x16_t vabdq_u8(uint8x16_t a, uint8x16_t b) //no abs for un
     return _mm_or_si128(difab, difba);
 }
 
-uint16x8_t vabdq_u16(uint16x8_t a, uint16x8_t b); // VABD.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vabdq_u16(uint16x8_t a, uint16x8_t b); // VABD.s16 q0,q0,q0
 _NEON2SSE_INLINE uint16x8_t vabdq_u16(uint16x8_t a, uint16x8_t b)
 {
     __m128i difab, difba;
@@ -5838,7 +5840,7 @@ _NEON2SSE_INLINE uint16x8_t vabdq_u16(uint16x8_t a, uint16x8_t b)
     return _mm_or_si128(difab, difba);
 }
 
-uint32x4_t vabdq_u32(uint32x4_t a, uint32x4_t b); // VABD.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vabdq_u32(uint32x4_t a, uint32x4_t b); // VABD.U32 q0,q0,q0
 _NEON2SSE_INLINE uint32x4_t vabdq_u32(uint32x4_t a, uint32x4_t b)
 {
     __m128i cmp, difab, difba;
@@ -5850,7 +5852,7 @@ _NEON2SSE_INLINE uint32x4_t vabdq_u32(uint32x4_t a, uint32x4_t b)
     return _mm_or_si128(difab, difba);
 }
 
-float32x4_t vabdq_f32(float32x4_t a, float32x4_t b); // VABD.F32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vabdq_f32(float32x4_t a, float32x4_t b); // VABD.F32 q0,q0,q0
 _NEON2SSE_INLINE float32x4_t vabdq_f32(float32x4_t a, float32x4_t b) // VABD.F32 q0,q0,q0
 {
     __m128i c1;
@@ -5862,7 +5864,7 @@ _NEON2SSE_INLINE float32x4_t vabdq_f32(float32x4_t a, float32x4_t b) // VABD.F32
 
 //************  Absolute difference - long **************************
 //********************************************************************
-int16x8_t vabdl_s8(int8x8_t a, int8x8_t b); // VABDL.S8 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vabdl_s8(int8x8_t a, int8x8_t b); // VABDL.S8 q0,d0,d0
 _NEON2SSE_INLINE int16x8_t vabdl_s8(int8x8_t a, int8x8_t b) // VABDL.S8 q0,d0,d0
 {
     __m128i a16, b16;
@@ -5872,7 +5874,7 @@ _NEON2SSE_INLINE int16x8_t vabdl_s8(int8x8_t a, int8x8_t b) // VABDL.S8 q0,d0,d0
 
 }
 
-int32x4_t vabdl_s16(int16x4_t a, int16x4_t b); // VABDL.S16 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vabdl_s16(int16x4_t a, int16x4_t b); // VABDL.S16 q0,d0,d0
 _NEON2SSE_INLINE int32x4_t vabdl_s16(int16x4_t a, int16x4_t b) // VABDL.S16 q0,d0,d0
 {
     __m128i a32, b32;
@@ -5881,7 +5883,7 @@ _NEON2SSE_INLINE int32x4_t vabdl_s16(int16x4_t a, int16x4_t b) // VABDL.S16 q0,d
     return vabdq_s32(a32, b32);
 }
 
-int64x2_t vabdl_s32(int32x2_t a, int32x2_t b); // VABDL.S32 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vabdl_s32(int32x2_t a, int32x2_t b); // VABDL.S32 q0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING (int64x2_t vabdl_s32(int32x2_t a, int32x2_t b),_NEON2SSE_REASON_SLOW_SERIAL)
 {
     //no optimal SIMD solution, serial looks faster
@@ -5893,7 +5895,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING (int64x2_t vabdl_s32(int32x2_t a,
     return _mm_load_si128((__m128i*)res);
 }
 
-uint16x8_t vabdl_u8(uint8x8_t a, uint8x8_t b); // VABDL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vabdl_u8(uint8x8_t a, uint8x8_t b); // VABDL.U8 q0,d0,d0
 _NEON2SSE_INLINE uint16x8_t vabdl_u8(uint8x8_t a, uint8x8_t b)
 {
     __m128i res;
@@ -5901,7 +5903,7 @@ _NEON2SSE_INLINE uint16x8_t vabdl_u8(uint8x8_t a, uint8x8_t b)
     return _mm_abs_epi16(res);
 }
 
-uint32x4_t vabdl_u16(uint16x4_t a, uint16x4_t b); // VABDL.s16 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vabdl_u16(uint16x4_t a, uint16x4_t b); // VABDL.s16 q0,d0,d0
 _NEON2SSE_INLINE uint32x4_t vabdl_u16(uint16x4_t a, uint16x4_t b)
 {
     __m128i res;
@@ -5909,7 +5911,7 @@ _NEON2SSE_INLINE uint32x4_t vabdl_u16(uint16x4_t a, uint16x4_t b)
     return _mm_abs_epi32(res);
 }
 
-uint64x2_t vabdl_u32(uint32x2_t a, uint32x2_t b); // VABDL.U32 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vabdl_u32(uint32x2_t a, uint32x2_t b); // VABDL.U32 q0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING (uint64x2_t vabdl_u32(uint32x2_t a, uint32x2_t b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     _NEON2SSE_ALIGN_16 uint64_t res[2];
@@ -5922,50 +5924,50 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING (uint64x2_t vabdl_u32(uint32x2_t
 
 //**********Absolute difference and accumulate: Vr[i] = Va[i] + | Vb[i] - Vc[i] | *************
 //*********************************************************************************************
-int8x8_t vaba_s8(int8x8_t a,  int8x8_t b, int8x8_t c); // VABA.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vaba_s8(int8x8_t a,  int8x8_t b, int8x8_t c); // VABA.S8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vaba_s8(int8x8_t a,  int8x8_t b, int8x8_t c)
 {
     int8x8_t res64;
     return64(vabaq_s8(_pM128i(a),_pM128i(b), _pM128i(c)));
 }
 
-int16x4_t vaba_s16(int16x4_t a,  int16x4_t b, int16x4_t c); // VABA.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vaba_s16(int16x4_t a,  int16x4_t b, int16x4_t c); // VABA.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vaba_s16(int16x4_t a,  int16x4_t b, int16x4_t c)
 {
     int16x4_t res64;
     return64(vabaq_s16(_pM128i(a), _pM128i(b), _pM128i(c)));
 }
 
-int32x2_t vaba_s32(int32x2_t a,  int32x2_t b, int32x2_t c); // VABA.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vaba_s32(int32x2_t a,  int32x2_t b, int32x2_t c); // VABA.S32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vaba_s32(int32x2_t a,  int32x2_t b, int32x2_t c)
 {
     int32x2_t res64;
     return64(vabaq_s32(_pM128i(a), _pM128i(b), _pM128i(c)));
 }
 
-uint8x8_t vaba_u8(uint8x8_t a,  uint8x8_t b, uint8x8_t c); // VABA.U8 d0,d0,d0
-_NEON2SSE_INLINE uint8x8_t vaba_u8(uint8x8_t a,  uint8x8_t b, uint8x8_t c) 
-{ 
-    int8x8_t res64; 
-    return64(vabaq_u8(_pM128i(a),_pM128i(b), _pM128i(c))); 
-} 
+_NEON2SSESTORAGE uint8x8_t vaba_u8(uint8x8_t a,  uint8x8_t b, uint8x8_t c); // VABA.U8 d0,d0,d0
+_NEON2SSE_INLINE uint8x8_t vaba_u8(uint8x8_t a,  uint8x8_t b, uint8x8_t c)
+{
+    int8x8_t res64;
+    return64(vabaq_u8(_pM128i(a),_pM128i(b), _pM128i(c)));
+}
 
 
-uint16x4_t vaba_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t c); // VABA.s16 d0,d0,d0
-_NEON2SSE_INLINE uint16x4_t vaba_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t c) 
-{ 
-    int16x4_t res64; 
-    return64(vabaq_u16(_pM128i(a), _pM128i(b), _pM128i(c))); 
-} 
+_NEON2SSESTORAGE uint16x4_t vaba_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t c); // VABA.s16 d0,d0,d0
+_NEON2SSE_INLINE uint16x4_t vaba_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t c)
+{
+    int16x4_t res64;
+    return64(vabaq_u16(_pM128i(a), _pM128i(b), _pM128i(c)));
+}
 
-uint32x2_t vaba_u32(uint32x2_t a,  uint32x2_t b, uint32x2_t c); // VABA.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vaba_u32(uint32x2_t a,  uint32x2_t b, uint32x2_t c); // VABA.U32 d0,d0,d0
 _NEON2SSE_INLINE uint32x2_t vaba_u32(uint32x2_t a,  uint32x2_t b, uint32x2_t c)
 {
     uint32x2_t res64;
     return64(vabaq_u32(_pM128i(a), _pM128i(b), _pM128i(c)));
 }
 
-int8x16_t vabaq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VABA.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vabaq_s8(int8x16_t a, int8x16_t b, int8x16_t c); // VABA.S8 q0,q0,q0
 _NEON2SSE_INLINE int8x16_t vabaq_s8(int8x16_t a, int8x16_t b, int8x16_t c) // VABA.S8 q0,q0,q0
 {
     int8x16_t sub;
@@ -5973,7 +5975,7 @@ _NEON2SSE_INLINE int8x16_t vabaq_s8(int8x16_t a, int8x16_t b, int8x16_t c) // VA
     return vaddq_s8( a, sub);
 }
 
-int16x8_t vabaq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VABA.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vabaq_s16(int16x8_t a, int16x8_t b, int16x8_t c); // VABA.S16 q0,q0,q0
 _NEON2SSE_INLINE int16x8_t vabaq_s16(int16x8_t a, int16x8_t b, int16x8_t c) // VABA.S16 q0,q0,q0
 {
     int16x8_t sub;
@@ -5981,7 +5983,7 @@ _NEON2SSE_INLINE int16x8_t vabaq_s16(int16x8_t a, int16x8_t b, int16x8_t c) // V
     return vaddq_s16( a, sub);
 }
 
-int32x4_t vabaq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VABA.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vabaq_s32(int32x4_t a, int32x4_t b, int32x4_t c); // VABA.S32 q0,q0,q0
 _NEON2SSE_INLINE int32x4_t vabaq_s32(int32x4_t a, int32x4_t b, int32x4_t c) // VABA.S32 q0,q0,q0
 {
     int32x4_t sub;
@@ -5989,7 +5991,7 @@ _NEON2SSE_INLINE int32x4_t vabaq_s32(int32x4_t a, int32x4_t b, int32x4_t c) // V
     return vaddq_s32( a, sub);
 }
 
-uint8x16_t vabaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VABA.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vabaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VABA.U8 q0,q0,q0
 _NEON2SSE_INLINE uint8x16_t vabaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c)
 {
     uint8x16_t sub;
@@ -5997,7 +5999,7 @@ _NEON2SSE_INLINE uint8x16_t vabaq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c)
     return vaddq_u8( a, sub);
 }
 
-uint16x8_t vabaq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VABA.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vabaq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VABA.s16 q0,q0,q0
 _NEON2SSE_INLINE uint16x8_t vabaq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c)
 {
     uint16x8_t sub;
@@ -6005,7 +6007,7 @@ _NEON2SSE_INLINE uint16x8_t vabaq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c)
     return vaddq_u16( a, sub);
 }
 
-uint32x4_t vabaq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VABA.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vabaq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VABA.U32 q0,q0,q0
 _NEON2SSE_INLINE uint32x4_t vabaq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c)
 {
     uint32x4_t sub;
@@ -6015,7 +6017,7 @@ _NEON2SSE_INLINE uint32x4_t vabaq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c)
 
 //************** Absolute difference and accumulate - long ********************************
 //*************************************************************************************
-int16x8_t vabal_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VABAL.S8 q0,d0,d0
+_NEON2SSESTORAGE int16x8_t vabal_s8(int16x8_t a, int8x8_t b, int8x8_t c); // VABAL.S8 q0,d0,d0
 _NEON2SSE_INLINE int16x8_t vabal_s8(int16x8_t a, int8x8_t b, int8x8_t c) // VABAL.S8 q0,d0,d0
 {
     __m128i b16, c16, res;
@@ -6025,7 +6027,7 @@ _NEON2SSE_INLINE int16x8_t vabal_s8(int16x8_t a, int8x8_t b, int8x8_t c) // VABA
     return _mm_add_epi16 (a, res);
 }
 
-int32x4_t vabal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VABAL.S16 q0,d0,d0
+_NEON2SSESTORAGE int32x4_t vabal_s16(int32x4_t a, int16x4_t b, int16x4_t c); // VABAL.S16 q0,d0,d0
 _NEON2SSE_INLINE int32x4_t vabal_s16(int32x4_t a, int16x4_t b, int16x4_t c) // VABAL.S16 q0,d0,d0
 {
     __m128i b32, c32, res;
@@ -6035,7 +6037,7 @@ _NEON2SSE_INLINE int32x4_t vabal_s16(int32x4_t a, int16x4_t b, int16x4_t c) // V
     return _mm_add_epi32 (a, res);
 }
 
-int64x2_t vabal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VABAL.S32 q0,d0,d0
+_NEON2SSESTORAGE int64x2_t vabal_s32(int64x2_t a, int32x2_t b, int32x2_t c); // VABAL.S32 q0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING (int64x2_t vabal_s32(int64x2_t a, int32x2_t b, int32x2_t c), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     __m128i res;
@@ -6043,7 +6045,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING (int64x2_t vabal_s32(int64x2_t a,
     return _mm_add_epi64(a, res);
 }
 
-uint16x8_t vabal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VABAL.U8 q0,d0,d0
+_NEON2SSESTORAGE uint16x8_t vabal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c); // VABAL.U8 q0,d0,d0
 _NEON2SSE_INLINE uint16x8_t vabal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c)
 {
     __m128i b16, c16, res;
@@ -6053,7 +6055,7 @@ _NEON2SSE_INLINE uint16x8_t vabal_u8(uint16x8_t a, uint8x8_t b, uint8x8_t c)
     return _mm_add_epi16 (a, res);
 }
 
-uint32x4_t vabal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VABAL.s16 q0,d0,d0
+_NEON2SSESTORAGE uint32x4_t vabal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c); // VABAL.s16 q0,d0,d0
 _NEON2SSE_INLINE uint32x4_t vabal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c)
 {
     __m128i b32, c32, res;
@@ -6063,7 +6065,7 @@ _NEON2SSE_INLINE uint32x4_t vabal_u16(uint32x4_t a, uint16x4_t b, uint16x4_t c)
     return _mm_add_epi32 (a, res);
 }
 
-uint64x2_t vabal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VABAL.U32 q0,d0,d0
+_NEON2SSESTORAGE uint64x2_t vabal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c); // VABAL.U32 q0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING (uint64x2_t vabal_u32(uint64x2_t a, uint32x2_t b, uint32x2_t c), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     __m128i res;
@@ -6076,7 +6078,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING (uint64x2_t vabal_u32(uint64x2_t
 //***********************************************************************************
 //************* Maximum:  vmax -> Vr[i] := (Va[i] >= Vb[i]) ? Va[i] : Vb[i]    *******
 //***********************************************************************************
-int8x8_t   vmax_s8(int8x8_t a, int8x8_t b); // VMAX.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t   vmax_s8(int8x8_t a, int8x8_t b); // VMAX.S8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t   vmax_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
@@ -6085,14 +6087,14 @@ _NEON2SSE_INLINE int8x8_t   vmax_s8(int8x8_t a, int8x8_t b)
     return64(res);
 }
 
-int16x4_t vmax_s16(int16x4_t a, int16x4_t b); // VMAX.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vmax_s16(int16x4_t a, int16x4_t b); // VMAX.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vmax_s16(int16x4_t a, int16x4_t b)
 {
     int16x4_t res64;
     return64(_mm_max_epi16(_pM128i(a),_pM128i(b)));
 }
 
-int32x2_t   vmax_s32(int32x2_t a, int32x2_t b); // VMAX.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t   vmax_s32(int32x2_t a, int32x2_t b); // VMAX.S32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t   vmax_s32(int32x2_t a, int32x2_t b)
 {
     int32x2_t res64;
@@ -6101,7 +6103,7 @@ _NEON2SSE_INLINE int32x2_t   vmax_s32(int32x2_t a, int32x2_t b)
     return64(res);
 }
 
-uint8x8_t vmax_u8(uint8x8_t a, uint8x8_t b); // VMAX.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vmax_u8(uint8x8_t a, uint8x8_t b); // VMAX.U8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vmax_u8(uint8x8_t a, uint8x8_t b)
 {
     uint8x8_t res64;
@@ -6109,7 +6111,7 @@ _NEON2SSE_INLINE uint8x8_t vmax_u8(uint8x8_t a, uint8x8_t b)
 }
 
 
-uint16x4_t vmax_u16(uint16x4_t a, uint16x4_t b); // VMAX.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vmax_u16(uint16x4_t a, uint16x4_t b); // VMAX.s16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vmax_u16(uint16x4_t a, uint16x4_t b)
 {
     uint16x4_t res64;
@@ -6117,7 +6119,7 @@ _NEON2SSE_INLINE uint16x4_t vmax_u16(uint16x4_t a, uint16x4_t b)
 }
 
 
-uint32x2_t   vmax_u32(uint32x2_t a, uint32x2_t b); // VMAX.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t   vmax_u32(uint32x2_t a, uint32x2_t b); // VMAX.U32 d0,d0,d0
 _NEON2SSE_INLINE uint32x2_t   vmax_u32(uint32x2_t a, uint32x2_t b)
 {
     uint32x2_t res64;
@@ -6126,7 +6128,7 @@ _NEON2SSE_INLINE uint32x2_t   vmax_u32(uint32x2_t a, uint32x2_t b)
     return64(res);
 }
 
-float32x2_t vmax_f32(float32x2_t a, float32x2_t b); // VMAX.F32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vmax_f32(float32x2_t a, float32x2_t b); // VMAX.F32 d0,d0,d0
 _NEON2SSE_INLINE float32x2_t vmax_f32(float32x2_t a, float32x2_t b)
 {
     //serial solution looks faster than  SIMD one
@@ -6136,36 +6138,36 @@ _NEON2SSE_INLINE float32x2_t vmax_f32(float32x2_t a, float32x2_t b)
     return res;
 }
 
-int8x16_t   vmaxq_s8(int8x16_t a, int8x16_t b); // VMAX.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t   vmaxq_s8(int8x16_t a, int8x16_t b); // VMAX.S8 q0,q0,q0
 #define vmaxq_s8 _MM_MAX_EPI8 //SSE4.1
 
-int16x8_t   vmaxq_s16(int16x8_t a, int16x8_t b); // VMAX.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t   vmaxq_s16(int16x8_t a, int16x8_t b); // VMAX.S16 q0,q0,q0
 #define vmaxq_s16 _mm_max_epi16
 
-int32x4_t   vmaxq_s32(int32x4_t a, int32x4_t b); // VMAX.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t   vmaxq_s32(int32x4_t a, int32x4_t b); // VMAX.S32 q0,q0,q0
 #define vmaxq_s32 _MM_MAX_EPI32 //SSE4.1
 
-uint8x16_t   vmaxq_u8(uint8x16_t a, uint8x16_t b); // VMAX.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t   vmaxq_u8(uint8x16_t a, uint8x16_t b); // VMAX.U8 q0,q0,q0
 #define vmaxq_u8 _mm_max_epu8
 
-uint16x8_t   vmaxq_u16(uint16x8_t a, uint16x8_t b); // VMAX.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   vmaxq_u16(uint16x8_t a, uint16x8_t b); // VMAX.s16 q0,q0,q0
 #define vmaxq_u16 _MM_MAX_EPU16 //SSE4.1
 
-uint32x4_t   vmaxq_u32(uint32x4_t a, uint32x4_t b); // VMAX.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t   vmaxq_u32(uint32x4_t a, uint32x4_t b); // VMAX.U32 q0,q0,q0
 #define vmaxq_u32 _MM_MAX_EPU32 //SSE4.1
 
 
-float32x4_t vmaxq_f32(float32x4_t a, float32x4_t b); // VMAX.F32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vmaxq_f32(float32x4_t a, float32x4_t b); // VMAX.F32 q0,q0,q0
 #define vmaxq_f32 _mm_max_ps
 
 
-float64x2_t vmaxq_f64(float64x2_t a, float64x2_t b); // VMAX.F64 q0,q0,q0
+_NEON2SSESTORAGE float64x2_t vmaxq_f64(float64x2_t a, float64x2_t b); // VMAX.F64 q0,q0,q0
 #define vmaxq_f64 _mm_max_pd
 
 
 //*************** Minimum: vmin -> Vr[i] := (Va[i] >= Vb[i]) ? Vb[i] : Va[i] ********************************
 //***********************************************************************************************************
-int8x8_t   vmin_s8(int8x8_t a, int8x8_t b); // VMIN.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t   vmin_s8(int8x8_t a, int8x8_t b); // VMIN.S8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t   vmin_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
@@ -6174,7 +6176,7 @@ _NEON2SSE_INLINE int8x8_t   vmin_s8(int8x8_t a, int8x8_t b)
     return64(res);
 }
 
-int16x4_t vmin_s16(int16x4_t a, int16x4_t b); // VMIN.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vmin_s16(int16x4_t a, int16x4_t b); // VMIN.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vmin_s16(int16x4_t a, int16x4_t b)
 {
     int16x4_t res64;
@@ -6182,7 +6184,7 @@ _NEON2SSE_INLINE int16x4_t vmin_s16(int16x4_t a, int16x4_t b)
 }
 
 
-int32x2_t   vmin_s32(int32x2_t a, int32x2_t b); // VMIN.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t   vmin_s32(int32x2_t a, int32x2_t b); // VMIN.S32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t   vmin_s32(int32x2_t a, int32x2_t b)
 {
     int32x2_t res64;
@@ -6191,7 +6193,7 @@ _NEON2SSE_INLINE int32x2_t   vmin_s32(int32x2_t a, int32x2_t b)
     return64(res);
 }
 
-uint8x8_t vmin_u8(uint8x8_t a, uint8x8_t b); // VMIN.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vmin_u8(uint8x8_t a, uint8x8_t b); // VMIN.U8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vmin_u8(uint8x8_t a, uint8x8_t b)
 {
     uint8x8_t res64;
@@ -6199,7 +6201,7 @@ _NEON2SSE_INLINE uint8x8_t vmin_u8(uint8x8_t a, uint8x8_t b)
 }
 
 
-uint16x4_t vmin_u16(uint16x4_t a, uint16x4_t b); // VMIN.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vmin_u16(uint16x4_t a, uint16x4_t b); // VMIN.s16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vmin_u16(uint16x4_t a, uint16x4_t b)
 {
     uint16x4_t res64;
@@ -6207,7 +6209,7 @@ _NEON2SSE_INLINE uint16x4_t vmin_u16(uint16x4_t a, uint16x4_t b)
 }
 
 
-uint32x2_t   vmin_u32(uint32x2_t a, uint32x2_t b); // VMIN.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t   vmin_u32(uint32x2_t a, uint32x2_t b); // VMIN.U32 d0,d0,d0
 _NEON2SSE_INLINE uint32x2_t   vmin_u32(uint32x2_t a, uint32x2_t b)
 {
     uint32x2_t res64;
@@ -6216,7 +6218,7 @@ _NEON2SSE_INLINE uint32x2_t   vmin_u32(uint32x2_t a, uint32x2_t b)
     return64(res);
 }
 
-float32x2_t vmin_f32(float32x2_t a, float32x2_t b); // VMIN.F32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vmin_f32(float32x2_t a, float32x2_t b); // VMIN.F32 d0,d0,d0
 _NEON2SSE_INLINE float32x2_t vmin_f32(float32x2_t a, float32x2_t b)
 {
     //serial solution looks faster than  SIMD one
@@ -6226,36 +6228,36 @@ _NEON2SSE_INLINE float32x2_t vmin_f32(float32x2_t a, float32x2_t b)
     return res;
 }
 
-int8x16_t   vminq_s8(int8x16_t a, int8x16_t b); // VMIN.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t   vminq_s8(int8x16_t a, int8x16_t b); // VMIN.S8 q0,q0,q0
 #define vminq_s8 _MM_MIN_EPI8 //SSE4.1
 
-int16x8_t   vminq_s16(int16x8_t a, int16x8_t b); // VMIN.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t   vminq_s16(int16x8_t a, int16x8_t b); // VMIN.S16 q0,q0,q0
 #define vminq_s16 _mm_min_epi16
 
-int32x4_t   vminq_s32(int32x4_t a, int32x4_t b); // VMIN.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t   vminq_s32(int32x4_t a, int32x4_t b); // VMIN.S32 q0,q0,q0
 #define vminq_s32 _MM_MIN_EPI32 //SSE4.1
 
-uint8x16_t   vminq_u8(uint8x16_t a, uint8x16_t b); // VMIN.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t   vminq_u8(uint8x16_t a, uint8x16_t b); // VMIN.U8 q0,q0,q0
 #define vminq_u8 _mm_min_epu8
 
-uint16x8_t   vminq_u16(uint16x8_t a, uint16x8_t b); // VMIN.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   vminq_u16(uint16x8_t a, uint16x8_t b); // VMIN.s16 q0,q0,q0
 #define vminq_u16 _MM_MIN_EPU16 //SSE4.1
 
-uint32x4_t   vminq_u32(uint32x4_t a, uint32x4_t b); // VMIN.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t   vminq_u32(uint32x4_t a, uint32x4_t b); // VMIN.U32 q0,q0,q0
 #define vminq_u32 _MM_MIN_EPU32 //SSE4.1
 
-float32x4_t vminq_f32(float32x4_t a, float32x4_t b); // VMIN.F32 q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vminq_f32(float32x4_t a, float32x4_t b); // VMIN.F32 q0,q0,q0
 #define vminq_f32 _mm_min_ps
 
 
-float64x2_t vminq_f64(float64x2_t a, float64x2_t b); // VMIN.F64 q0,q0,q0
+_NEON2SSESTORAGE float64x2_t vminq_f64(float64x2_t a, float64x2_t b); // VMIN.F64 q0,q0,q0
 #define vminq_f64 _mm_min_pd
 
 
 //*************  Pairwise addition operations. **************************************
 //************************************************************************************
 //Pairwise add - adds adjacent pairs of elements of two vectors, and places the results in the destination vector
-int8x8_t vpadd_s8(int8x8_t a, int8x8_t b); // VPADD.I8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vpadd_s8(int8x8_t a, int8x8_t b); // VPADD.I8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vpadd_s8(int8x8_t a, int8x8_t b) // VPADD.I8 d0,d0,d0
 {
     //no 8 bit hadd in IA32, need to go to 16 bit and then pack
@@ -6268,7 +6270,7 @@ _NEON2SSE_INLINE int8x8_t vpadd_s8(int8x8_t a, int8x8_t b) // VPADD.I8 d0,d0,d0
     return64(res);
 }
 
-int16x4_t   vpadd_s16(int16x4_t a, int16x4_t b); // VPADD.I16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t   vpadd_s16(int16x4_t a, int16x4_t b); // VPADD.I16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t   vpadd_s16(int16x4_t a, int16x4_t b)
 {
     int16x4_t res64;
@@ -6279,7 +6281,7 @@ _NEON2SSE_INLINE int16x4_t   vpadd_s16(int16x4_t a, int16x4_t b)
 }
 
 
-int32x2_t   vpadd_s32(int32x2_t a, int32x2_t b); // VPADD.I32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t   vpadd_s32(int32x2_t a, int32x2_t b); // VPADD.I32 d0,d0,d0
 _NEON2SSE_INLINE int32x2_t   vpadd_s32(int32x2_t a, int32x2_t b)
 {
     int32x2_t res64;
@@ -6290,7 +6292,7 @@ _NEON2SSE_INLINE int32x2_t   vpadd_s32(int32x2_t a, int32x2_t b)
 }
 
 
-uint8x8_t vpadd_u8(uint8x8_t a, uint8x8_t b); // VPADD.I8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vpadd_u8(uint8x8_t a, uint8x8_t b); // VPADD.I8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vpadd_u8(uint8x8_t a, uint8x8_t b) // VPADD.I8 d0,d0,d0
 {
     //  no 8 bit hadd in IA32, need to go to 16 bit and then pack
@@ -6306,7 +6308,7 @@ _NEON2SSE_INLINE uint8x8_t vpadd_u8(uint8x8_t a, uint8x8_t b) // VPADD.I8 d0,d0,
     return64(res);
 }
 
-uint16x4_t vpadd_u16(uint16x4_t a, uint16x4_t b); // VPADD.I16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vpadd_u16(uint16x4_t a, uint16x4_t b); // VPADD.I16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vpadd_u16(uint16x4_t a, uint16x4_t b) // VPADD.I16 d0,d0,d0
 {
     // solution may be not optimal, serial execution may be faster
@@ -6323,7 +6325,7 @@ _NEON2SSE_INLINE uint16x4_t vpadd_u16(uint16x4_t a, uint16x4_t b) // VPADD.I16 d
     return64(res);
 }
 
-uint32x2_t vpadd_u32(uint32x2_t a, uint32x2_t b); // VPADD.I32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vpadd_u32(uint32x2_t a, uint32x2_t b); // VPADD.I32 d0,d0,d0
 _NEON2SSE_INLINE uint32x2_t vpadd_u32(uint32x2_t a, uint32x2_t b) //serial may be faster
 {
     //hadd doesn't work for unsigned values
@@ -6336,7 +6338,7 @@ _NEON2SSE_INLINE uint32x2_t vpadd_u32(uint32x2_t a, uint32x2_t b) //serial may b
     return64(res);
 }
 
-float32x2_t vpadd_f32(float32x2_t a, float32x2_t b); // VPADD.F32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vpadd_f32(float32x2_t a, float32x2_t b); // VPADD.F32 d0,d0,d0
 _NEON2SSE_INLINE float32x2_t vpadd_f32(float32x2_t a, float32x2_t b)
 {
     __m128 hadd128;
@@ -6353,7 +6355,7 @@ _NEON2SSE_INLINE float32x2_t vpadd_f32(float32x2_t a, float32x2_t b)
 //Adds adjacent pairs of elements of a vector,sign or zero extends the results to twice their original width,
 // and places the final results in the destination vector.
 
-int16x4_t vpaddl_s8(int8x8_t a); // VPADDL.S8 d0,d0
+_NEON2SSESTORAGE int16x4_t vpaddl_s8(int8x8_t a); // VPADDL.S8 d0,d0
 _NEON2SSE_INLINE int16x4_t vpaddl_s8(int8x8_t a) // VPADDL.S8 d0,d0
 {
     //no 8 bit hadd in IA32, need to go to 16 bit anyway
@@ -6364,7 +6366,7 @@ _NEON2SSE_INLINE int16x4_t vpaddl_s8(int8x8_t a) // VPADDL.S8 d0,d0
     return64(a16);
 }
 
-int32x2_t vpaddl_s16(int16x4_t a); // VPADDL.S16 d0,d0
+_NEON2SSESTORAGE int32x2_t vpaddl_s16(int16x4_t a); // VPADDL.S16 d0,d0
 _NEON2SSE_INLINE int32x2_t vpaddl_s16(int16x4_t a) // VPADDL.S16 d0,d0
 {
     // solution may be not optimal, serial execution may be faster
@@ -6375,7 +6377,7 @@ _NEON2SSE_INLINE int32x2_t vpaddl_s16(int16x4_t a) // VPADDL.S16 d0,d0
     return64(r32_1);
 }
 
-int64x1_t vpaddl_s32(int32x2_t a); // VPADDL.S32 d0,d0
+_NEON2SSESTORAGE int64x1_t vpaddl_s32(int32x2_t a); // VPADDL.S32 d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vpaddl_s32(int32x2_t a), _NEON2SSE_REASON_SLOW_SERIAL) //serial solution looks faster
 {
     int64x1_t res;
@@ -6383,7 +6385,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vpaddl_s32(int32x2_t a)
     return res;
 }
 
-uint16x4_t vpaddl_u8(uint8x8_t a); // VPADDL.U8 d0,d0
+_NEON2SSESTORAGE uint16x4_t vpaddl_u8(uint8x8_t a); // VPADDL.U8 d0,d0
 _NEON2SSE_INLINE uint16x4_t vpaddl_u8(uint8x8_t a) // VPADDL.U8 d0,d0
 {
     //  no 8 bit hadd in IA32, need to go to 16 bit
@@ -6395,7 +6397,7 @@ _NEON2SSE_INLINE uint16x4_t vpaddl_u8(uint8x8_t a) // VPADDL.U8 d0,d0
     return64(a16);
 }
 
-uint32x2_t vpaddl_u16(uint16x4_t a); // VPADDL.s16 d0,d0
+_NEON2SSESTORAGE uint32x2_t vpaddl_u16(uint16x4_t a); // VPADDL.s16 d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vpaddl_u16(uint16x4_t a),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //serial solution looks faster than a SIMD one
@@ -6405,7 +6407,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vpaddl_u16(uint16x4_t
     return res;
 }
 
-uint64x1_t vpaddl_u32(uint32x2_t a); // VPADDL.U32 d0,d0
+_NEON2SSESTORAGE uint64x1_t vpaddl_u32(uint32x2_t a); // VPADDL.U32 d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vpaddl_u32(uint32x2_t a), _NEON2SSE_REASON_SLOW_SERIAL) //serial solution looks faster
 {
     uint64x1_t res;
@@ -6413,7 +6415,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vpaddl_u32(uint32x2_t
     return res;
 }
 
-int16x8_t vpaddlq_s8(int8x16_t a); // VPADDL.S8 q0,q0
+_NEON2SSESTORAGE int16x8_t vpaddlq_s8(int8x16_t a); // VPADDL.S8 q0,q0
 _NEON2SSE_INLINE int16x8_t vpaddlq_s8(int8x16_t a) // VPADDL.S8 q0,q0
 {
     //no 8 bit hadd in IA32, need to go to 16 bit
@@ -6425,7 +6427,7 @@ _NEON2SSE_INLINE int16x8_t vpaddlq_s8(int8x16_t a) // VPADDL.S8 q0,q0
     return _mm_hadd_epi16 (r16_1, r16_2);
 }
 
-int32x4_t vpaddlq_s16(int16x8_t a); // VPADDL.S16 q0,q0
+_NEON2SSESTORAGE int32x4_t vpaddlq_s16(int16x8_t a); // VPADDL.S16 q0,q0
 _NEON2SSE_INLINE int32x4_t vpaddlq_s16(int16x8_t a) // VPADDL.S16 q0,q0
 {
     //no 8 bit hadd in IA32, need to go to 16 bit
@@ -6437,7 +6439,7 @@ _NEON2SSE_INLINE int32x4_t vpaddlq_s16(int16x8_t a) // VPADDL.S16 q0,q0
     return _mm_hadd_epi32 (r32_1, r32_2);
 }
 
-int64x2_t vpaddlq_s32(int32x4_t a); // VPADDL.S32 q0,q0
+_NEON2SSESTORAGE int64x2_t vpaddlq_s32(int32x4_t a); // VPADDL.S32 q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vpaddlq_s32(int32x4_t a), _NEON2SSE_REASON_SLOW_SERIAL) // VPADDL.S32 q0,q0
 {
     _NEON2SSE_ALIGN_16 int32_t atmp[4];
@@ -6448,7 +6450,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vpaddlq_s32(int32x4_t a
     return _mm_load_si128((__m128i*)res);
 }
 
-uint16x8_t vpaddlq_u8(uint8x16_t a); // VPADDL.U8 q0,q0
+_NEON2SSESTORAGE uint16x8_t vpaddlq_u8(uint8x16_t a); // VPADDL.U8 q0,q0
 _NEON2SSE_INLINE uint16x8_t vpaddlq_u8(uint8x16_t a) // VPADDL.U8 q0,q0
 {
     //no 8 bit hadd in IA32, need to go to 16 bit
@@ -6460,7 +6462,7 @@ _NEON2SSE_INLINE uint16x8_t vpaddlq_u8(uint8x16_t a) // VPADDL.U8 q0,q0
     return _mm_hadd_epi16 (r16_1, r16_2);
 }
 
-uint32x4_t vpaddlq_u16(uint16x8_t a); // VPADDL.s16 q0,q0
+_NEON2SSESTORAGE uint32x4_t vpaddlq_u16(uint16x8_t a); // VPADDL.s16 q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vpaddlq_u16(uint16x8_t a),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //serial solution looks faster than a SIMD one
@@ -6474,7 +6476,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vpaddlq_u16(uint16x8_t
     return _mm_load_si128((__m128i*)res);
 }
 
-uint64x2_t vpaddlq_u32(uint32x4_t a); // VPADDL.U32 q0,q0
+_NEON2SSESTORAGE uint64x2_t vpaddlq_u32(uint32x4_t a); // VPADDL.U32 q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vpaddlq_u32(uint32x4_t a), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     _NEON2SSE_ALIGN_16 uint32_t atmp[4];
@@ -6489,14 +6491,14 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vpaddlq_u32(uint32x4_t
 //****************************************************************************************
 //VPADAL (Vector Pairwise Add and Accumulate Long) adds adjacent pairs of elements of a vector,
 // and accumulates the  values of the results into the elements of the destination (wide) vector
-int16x4_t vpadal_s8(int16x4_t a,  int8x8_t b); // VPADAL.S8 d0,d0
+_NEON2SSESTORAGE int16x4_t vpadal_s8(int16x4_t a,  int8x8_t b); // VPADAL.S8 d0,d0
 _NEON2SSE_INLINE int16x4_t vpadal_s8(int16x4_t a,  int8x8_t b)
 {
     int16x4_t res64;
     return64(vpadalq_s8(_pM128i(a), _pM128i(b)));
 }
 
-int32x2_t vpadal_s16(int32x2_t a,  int16x4_t b); // VPADAL.S16 d0,d0
+_NEON2SSESTORAGE int32x2_t vpadal_s16(int32x2_t a,  int16x4_t b); // VPADAL.S16 d0,d0
 _NEON2SSE_INLINE int32x2_t vpadal_s16(int32x2_t a,  int16x4_t b)
 {
     int32x2_t res64;
@@ -6504,7 +6506,7 @@ _NEON2SSE_INLINE int32x2_t vpadal_s16(int32x2_t a,  int16x4_t b)
 }
 
 
-int64x1_t vpadal_s32(int64x1_t a, int32x2_t b); // VPADAL.S32 d0,d0
+_NEON2SSESTORAGE int64x1_t vpadal_s32(int64x1_t a, int32x2_t b); // VPADAL.S32 d0,d0
 _NEON2SSE_INLINE int64x1_t vpadal_s32(int64x1_t a, int32x2_t b)
 {
     int64x1_t res;
@@ -6512,7 +6514,7 @@ _NEON2SSE_INLINE int64x1_t vpadal_s32(int64x1_t a, int32x2_t b)
     return res;
 }
 
-uint16x4_t vpadal_u8(uint16x4_t a,  uint8x8_t b); // VPADAL.U8 d0,d0
+_NEON2SSESTORAGE uint16x4_t vpadal_u8(uint16x4_t a,  uint8x8_t b); // VPADAL.U8 d0,d0
 _NEON2SSE_INLINE uint16x4_t vpadal_u8(uint16x4_t a,  uint8x8_t b)
 {
     uint16x4_t res64;
@@ -6520,14 +6522,14 @@ _NEON2SSE_INLINE uint16x4_t vpadal_u8(uint16x4_t a,  uint8x8_t b)
 }
 
 
-uint32x2_t vpadal_u16(uint32x2_t a,  uint16x4_t b); // VPADAL.s16 d0,d0
+_NEON2SSESTORAGE uint32x2_t vpadal_u16(uint32x2_t a,  uint16x4_t b); // VPADAL.s16 d0,d0
 _NEON2SSE_INLINE uint32x2_t vpadal_u16(uint32x2_t a,  uint16x4_t b)
 {
     uint32x2_t res64;
     return64(vpadalq_u16(_pM128i(a), _pM128i(b)));
 }
 
-uint64x1_t vpadal_u32(uint64x1_t a, uint32x2_t b); // VPADAL.U32 d0,d0
+_NEON2SSESTORAGE uint64x1_t vpadal_u32(uint64x1_t a, uint32x2_t b); // VPADAL.U32 d0,d0
 _NEON2SSE_INLINE uint64x1_t vpadal_u32(uint64x1_t a, uint32x2_t b)
 {
     uint64x1_t res;
@@ -6535,7 +6537,7 @@ _NEON2SSE_INLINE uint64x1_t vpadal_u32(uint64x1_t a, uint32x2_t b)
     return res;
 }
 
-int16x8_t vpadalq_s8(int16x8_t a, int8x16_t b); // VPADAL.S8 q0,q0
+_NEON2SSESTORAGE int16x8_t vpadalq_s8(int16x8_t a, int8x16_t b); // VPADAL.S8 q0,q0
 _NEON2SSE_INLINE int16x8_t vpadalq_s8(int16x8_t a, int8x16_t b) // VPADAL.S8 q0,q0
 {
     int16x8_t pad;
@@ -6543,7 +6545,7 @@ _NEON2SSE_INLINE int16x8_t vpadalq_s8(int16x8_t a, int8x16_t b) // VPADAL.S8 q0,
     return _mm_add_epi16 (a, pad);
 }
 
-int32x4_t vpadalq_s16(int32x4_t a, int16x8_t b); // VPADAL.S16 q0,q0
+_NEON2SSESTORAGE int32x4_t vpadalq_s16(int32x4_t a, int16x8_t b); // VPADAL.S16 q0,q0
 _NEON2SSE_INLINE int32x4_t vpadalq_s16(int32x4_t a, int16x8_t b) // VPADAL.S16 q0,q0
 {
     int32x4_t pad;
@@ -6551,7 +6553,7 @@ _NEON2SSE_INLINE int32x4_t vpadalq_s16(int32x4_t a, int16x8_t b) // VPADAL.S16 q
     return _mm_add_epi32(a, pad);
 }
 
-int64x2_t vpadalq_s32(int64x2_t a, int32x4_t b); // VPADAL.S32 q0,q0
+_NEON2SSESTORAGE int64x2_t vpadalq_s32(int64x2_t a, int32x4_t b); // VPADAL.S32 q0,q0
 _NEON2SSE_INLINE int64x2_t vpadalq_s32(int64x2_t a, int32x4_t b)
 {
     int64x2_t pad;
@@ -6559,7 +6561,7 @@ _NEON2SSE_INLINE int64x2_t vpadalq_s32(int64x2_t a, int32x4_t b)
     return _mm_add_epi64 (a, pad);
 }
 
-uint16x8_t vpadalq_u8(uint16x8_t a, uint8x16_t b); // VPADAL.U8 q0,q0
+_NEON2SSESTORAGE uint16x8_t vpadalq_u8(uint16x8_t a, uint8x16_t b); // VPADAL.U8 q0,q0
 _NEON2SSE_INLINE uint16x8_t vpadalq_u8(uint16x8_t a, uint8x16_t b) // VPADAL.U8 q0,q0
 {
     uint16x8_t pad;
@@ -6567,7 +6569,7 @@ _NEON2SSE_INLINE uint16x8_t vpadalq_u8(uint16x8_t a, uint8x16_t b) // VPADAL.U8
     return _mm_add_epi16 (a, pad);
 }
 
-uint32x4_t vpadalq_u16(uint32x4_t a, uint16x8_t b); // VPADAL.s16 q0,q0
+_NEON2SSESTORAGE uint32x4_t vpadalq_u16(uint32x4_t a, uint16x8_t b); // VPADAL.s16 q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vpadalq_u16(uint32x4_t a, uint16x8_t b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     uint32x4_t pad;
@@ -6575,7 +6577,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vpadalq_u16(uint32x4_t
     return _mm_add_epi32(a, pad);
 } //no optimal SIMD solution, serial is faster
 
-uint64x2_t vpadalq_u32(uint64x2_t a, uint32x4_t b); // VPADAL.U32 q0,q0
+_NEON2SSESTORAGE uint64x2_t vpadalq_u32(uint64x2_t a, uint32x4_t b); // VPADAL.U32 q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vpadalq_u32(uint64x2_t a, uint32x4_t b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //no optimal SIMD solution, serial is faster
@@ -6589,13 +6591,13 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vpadalq_u32(uint64x2_t
 //VPMAX (Vector Pairwise Maximum) compares adjacent pairs of elements in two vectors,
 //and copies the larger of each pair into the corresponding element in the destination
 //    no corresponding functionality in IA32 SIMD, so we need to do the vertical comparison
-int8x8_t vpmax_s8(int8x8_t a, int8x8_t b); // VPMAX.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vpmax_s8(int8x8_t a, int8x8_t b); // VPMAX.S8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vpmax_s8(int8x8_t a, int8x8_t b) // VPMAX.S8 d0,d0,d0
 {
     int8x8_t res64;
     __m128i ab, ab1, max;
-    _NEON2SSE_ALIGN_16 uint8_t mask8_sab[16] = { 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14};
-    _NEON2SSE_ALIGN_16 uint8_t mask8_odd[16] = { 1, 3,  5,  7, 9, 11, 13, 15, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask8_sab[16] = { 1, 0, 3, 2, 5,  4,  7,  6,    9,    8,   11,   10,   13,   12,   15,   14};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask8_odd[16] = { 1, 3, 5, 7, 9, 11, 13, 15, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff};
     ab = _mm_unpacklo_epi64 ( _pM128i(a), _pM128i(b)); //ab
     ab1 = _mm_shuffle_epi8 (ab, *(__m128i*) mask8_sab); //horisontal pairs swap for vertical max finding
     max = _MM_MAX_EPI8 (ab, ab1); // SSE4.1
@@ -6603,13 +6605,13 @@ _NEON2SSE_INLINE int8x8_t vpmax_s8(int8x8_t a, int8x8_t b) // VPMAX.S8 d0,d0,d0
     return64(max); //we need 64 bits only
 }
 
-int16x4_t vpmax_s16(int16x4_t a, int16x4_t b); // VPMAX.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vpmax_s16(int16x4_t a, int16x4_t b); // VPMAX.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vpmax_s16(int16x4_t a, int16x4_t b) // VPMAX.S16 d0,d0,d0
 {
     //solution may be not optimal compared with the serial one
     int16x4_t res64;
     __m128i ab, ab1, max;
-    _NEON2SSE_ALIGN_16 int8_t mask16_sab[16] = { 2, 3, 0, 1, 6, 7, 4, 5, 10, 11, 8, 9, 14, 15, 12, 13}; //each chars pair is considerd to be 16 bit number
+    _NEON2SSE_ALIGN_16 static const int8_t mask16_sab[16] = { 2, 3, 0, 1, 6, 7, 4, 5, 10, 11, 8, 9, 14, 15, 12, 13}; //each chars pair is considerd to be 16 bit number
     ab = _mm_unpacklo_epi64 ( _pM128i(a),  _pM128i(b)); //ab
     ab1 = _mm_shuffle_epi8 (ab, *(__m128i*) mask16_sab); //horisontal pairs swap for vertical max finding, use 8bit fn and the corresponding mask
     max = _mm_max_epi16 (ab, ab1);
@@ -6617,7 +6619,7 @@ _NEON2SSE_INLINE int16x4_t vpmax_s16(int16x4_t a, int16x4_t b) // VPMAX.S16 d0,d
     return64(max);
 }
 
-int32x2_t vpmax_s32(int32x2_t a, int32x2_t b); // VPMAX.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vpmax_s32(int32x2_t a, int32x2_t b); // VPMAX.S32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vpmax_s32(int32x2_t a, int32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //serial solution looks faster than SIMD one
@@ -6627,13 +6629,13 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vpmax_s32(int32x2_t a,
     return res;
 }
 
-uint8x8_t vpmax_u8(uint8x8_t a, uint8x8_t b); // VPMAX.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vpmax_u8(uint8x8_t a, uint8x8_t b); // VPMAX.U8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vpmax_u8(uint8x8_t a, uint8x8_t b) // VPMAX.U8 d0,d0,d0
 {
     uint8x8_t res64;
     __m128i ab, ab1, max;
-    _NEON2SSE_ALIGN_16 int8_t mask8_sab[16] = { 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14};
-    _NEON2SSE_ALIGN_16 uint8_t mask8_odd[16] = { 1, 3,  5,  7, 9, 11, 13, 15, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff};
+    _NEON2SSE_ALIGN_16 static const int8_t mask8_sab[16] = { 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask8_odd[16] = { 1, 3,  5,  7, 9, 11, 13, 15, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff};
     ab = _mm_unpacklo_epi64 (_pM128i(a), _pM128i(b)); //ab
     ab1 = _mm_shuffle_epi8 (ab, *(__m128i*) mask8_sab); //horisontal pairs swap for vertical max finding
     max = _mm_max_epu8 (ab, ab1); // SSE4.1
@@ -6641,13 +6643,13 @@ _NEON2SSE_INLINE uint8x8_t vpmax_u8(uint8x8_t a, uint8x8_t b) // VPMAX.U8 d0,d0,
     return64(max);
 }
 
-uint16x4_t vpmax_u16(uint16x4_t a, uint16x4_t b); // VPMAX.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vpmax_u16(uint16x4_t a, uint16x4_t b); // VPMAX.s16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vpmax_u16(uint16x4_t a, uint16x4_t b) // VPMAX.s16 d0,d0,d0
 {
     //solution may be not optimal compared with the serial one
     uint16x4_t res64;
     __m128i ab, ab1, max;
-    _NEON2SSE_ALIGN_16 uint8_t mask16_sab[16] = { 2, 3, 0, 1, 6, 7, 4, 5, 10, 11, 8, 9, 14, 15, 12, 13}; //each chars pair is considerd to be 16 bit number
+    _NEON2SSE_ALIGN_16 static const uint8_t mask16_sab[16] = { 2, 3, 0, 1, 6, 7, 4, 5, 10, 11, 8, 9, 14, 15, 12, 13}; //each chars pair is considerd to be 16 bit number
     ab = _mm_unpacklo_epi64 ( _pM128i(a), _pM128i(b)); //ab
     ab1 = _mm_shuffle_epi8 (ab, *(__m128i*) mask16_sab); //horisontal pairs swap for vertical max finding, use 8bit fn and the corresponding mask
     max = _MM_MAX_EPU16 (ab, ab1);
@@ -6655,7 +6657,7 @@ _NEON2SSE_INLINE uint16x4_t vpmax_u16(uint16x4_t a, uint16x4_t b) // VPMAX.s16 d
     return64(max);
 }
 
-uint32x2_t vpmax_u32(uint32x2_t a, uint32x2_t b); // VPMAX.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vpmax_u32(uint32x2_t a, uint32x2_t b); // VPMAX.U32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vpmax_u32(uint32x2_t a, uint32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //serial solution looks faster than SIMD one
@@ -6663,9 +6665,9 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vpmax_u32(uint32x2_t a
     res.m64_u32[0] = (a.m64_u32[0] < a.m64_u32[1]) ? a.m64_u32[1] : a.m64_u32[0];
     res.m64_u32[1] = (b.m64_u32[0] < b.m64_u32[1]) ? b.m64_u32[1] : b.m64_u32[0];
     return res;
-} 
+}
 
-float32x2_t vpmax_f32(float32x2_t a, float32x2_t b); // VPMAX.F32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vpmax_f32(float32x2_t a, float32x2_t b); // VPMAX.F32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(float32x2_t vpmax_f32(float32x2_t a, float32x2_t b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //serial solution looks faster than  SIMD one
@@ -6678,13 +6680,13 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(float32x2_t vpmax_f32(float32x2_t
 // ***************** Folding minimum  ****************************
 // **************************************************************
 //vpmin -> takes minimum of adjacent pairs
-int8x8_t vpmin_s8(int8x8_t a, int8x8_t b); // VPMIN.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vpmin_s8(int8x8_t a, int8x8_t b); // VPMIN.S8 d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vpmin_s8(int8x8_t a, int8x8_t b) // VPMIN.S8 d0,d0,d0
 {
     int8x8_t res64;
     __m128i ab, ab1, min;
-    _NEON2SSE_ALIGN_16 uint8_t mask8_sab[16] = { 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14};
-    _NEON2SSE_ALIGN_16 uint8_t mask8_odd[16] = { 1, 3,  5,  7, 9, 11, 13, 15, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask8_sab[16] = { 1, 0, 3, 2, 5,  4,  7,  6,    9,    8,   11,   10,   13,   12,   15,   14};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask8_odd[16] = { 1, 3, 5, 7, 9, 11, 13, 15, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff};
     ab = _mm_unpacklo_epi64 ( _pM128i(a), _pM128i(b)); //ab
     ab1 = _mm_shuffle_epi8 (ab, *(__m128i*) mask8_sab); //horisontal pairs swap for vertical min finding
     min =  _MM_MIN_EPI8 (ab, ab1); // SSE4.1
@@ -6692,13 +6694,13 @@ _NEON2SSE_INLINE int8x8_t vpmin_s8(int8x8_t a, int8x8_t b) // VPMIN.S8 d0,d0,d0
     return64(min);
 }
 
-int16x4_t vpmin_s16(int16x4_t a, int16x4_t b); // VPMIN.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vpmin_s16(int16x4_t a, int16x4_t b); // VPMIN.S16 d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vpmin_s16(int16x4_t a, int16x4_t b) // VPMIN.S16 d0,d0,d0
 {
     //solution may be not optimal compared with the serial one
     int16x4_t res64;
     __m128i ab, ab1, min;
-    _NEON2SSE_ALIGN_16 int8_t mask16_sab[16] = { 2, 3, 0, 1, 6, 7, 4, 5, 10, 11, 8, 9, 14, 15, 12, 13}; //each chars pair is considerd to be 16 bit number
+    _NEON2SSE_ALIGN_16 static const int8_t mask16_sab[16] = { 2, 3, 0, 1, 6, 7, 4, 5, 10, 11, 8, 9, 14, 15, 12, 13}; //each chars pair is considerd to be 16 bit number
     ab = _mm_unpacklo_epi64 (  _pM128i(a),  _pM128i(b)); //ab
     ab1 = _mm_shuffle_epi8 (ab, *(__m128i*) mask16_sab); //horisontal pairs swap for vertical max finding, use 8bit fn and the corresponding mask
     min = _mm_min_epi16 (ab, ab1);
@@ -6706,7 +6708,7 @@ _NEON2SSE_INLINE int16x4_t vpmin_s16(int16x4_t a, int16x4_t b) // VPMIN.S16 d0,d
     return64(min);
 }
 
-int32x2_t vpmin_s32(int32x2_t a, int32x2_t b); // VPMIN.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vpmin_s32(int32x2_t a, int32x2_t b); // VPMIN.S32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vpmin_s32(int32x2_t a, int32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //serial solution looks faster than SIMD one
@@ -6716,13 +6718,13 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vpmin_s32(int32x2_t a,
     return res;
 }
 
-uint8x8_t vpmin_u8(uint8x8_t a, uint8x8_t b); // VPMIN.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vpmin_u8(uint8x8_t a, uint8x8_t b); // VPMIN.U8 d0,d0,d0
 _NEON2SSE_INLINE uint8x8_t vpmin_u8(uint8x8_t a, uint8x8_t b) // VPMIN.U8 d0,d0,d0
 {
     uint8x8_t res64;
     __m128i ab, ab1, min;
-    _NEON2SSE_ALIGN_16 uint8_t mask8_sab[16] = { 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14};
-    _NEON2SSE_ALIGN_16 uint8_t mask8_odd[16] = { 1, 3,  5,  7, 9, 11, 13, 15, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask8_sab[16] = { 1, 0, 3, 2, 5,  4,  7,  6,    9,    8,   11,   10,   13,   12,   15,   14};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask8_odd[16] = { 1, 3, 5, 7, 9, 11, 13, 15, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff};
     ab = _mm_unpacklo_epi64 (  _pM128i(a),  _pM128i(b)); //ab
     ab1 = _mm_shuffle_epi8 (ab, *(__m128i*) mask8_sab); //horisontal pairs swap for vertical max finding
     min = _mm_min_epu8 (ab, ab1); // SSE4.1
@@ -6730,13 +6732,13 @@ _NEON2SSE_INLINE uint8x8_t vpmin_u8(uint8x8_t a, uint8x8_t b) // VPMIN.U8 d0,d0,
     return64(min);
 }
 
-uint16x4_t vpmin_u16(uint16x4_t a, uint16x4_t b); // VPMIN.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vpmin_u16(uint16x4_t a, uint16x4_t b); // VPMIN.s16 d0,d0,d0
 _NEON2SSE_INLINE uint16x4_t vpmin_u16(uint16x4_t a, uint16x4_t b) // VPMIN.s16 d0,d0,d0
 {
     //solution may be not optimal compared with the serial one
     uint16x4_t res64;
     __m128i ab, ab1, min;
-    _NEON2SSE_ALIGN_16 uint8_t mask16_sab[16] = { 2, 3, 0, 1, 6, 7, 4, 5, 10, 11, 8, 9, 14, 15, 12, 13}; //each chars pair is considerd to be 16 bit number
+    _NEON2SSE_ALIGN_16 static const uint8_t mask16_sab[16] = { 2, 3, 0, 1, 6, 7, 4, 5, 10, 11, 8, 9, 14, 15, 12, 13}; //each chars pair is considerd to be 16 bit number
     ab = _mm_unpacklo_epi64 ( _pM128i(a),  _pM128i(b)); //ab
     ab1 = _mm_shuffle_epi8 (ab, *(__m128i*) mask16_sab); //horisontal pairs swap for vertical min finding, use 8bit fn and the corresponding mask
     min = _MM_MIN_EPU16 (ab, ab1);
@@ -6744,7 +6746,7 @@ _NEON2SSE_INLINE uint16x4_t vpmin_u16(uint16x4_t a, uint16x4_t b) // VPMIN.s16 d
     return64(min);
 }
 
-uint32x2_t vpmin_u32(uint32x2_t a, uint32x2_t b); // VPMIN.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vpmin_u32(uint32x2_t a, uint32x2_t b); // VPMIN.U32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vpmin_u32(uint32x2_t a, uint32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //serial solution looks faster than SIMD one
@@ -6754,7 +6756,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vpmin_u32(uint32x2_t a
     return res;
 }
 
-float32x2_t vpmin_f32(float32x2_t a, float32x2_t b); // VPMIN.F32 d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vpmin_f32(float32x2_t a, float32x2_t b); // VPMIN.F32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(float32x2_t vpmin_f32(float32x2_t a, float32x2_t b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //serial solution looks faster than SIMD one
@@ -6769,7 +6771,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(float32x2_t vpmin_f32(float32x2_t
 //***************************************************************
 //****************** Reciprocal estimate *******************************
 //the ARM NEON and x86 SIMD results may be slightly different
-float32x2_t vrecpe_f32(float32x2_t a); // VRECPE.F32 d0,d0
+_NEON2SSESTORAGE float32x2_t vrecpe_f32(float32x2_t a); // VRECPE.F32 d0,d0
 _NEON2SSE_INLINE float32x2_t vrecpe_f32(float32x2_t a) //use low 64 bits
 {
     float32x4_t res;
@@ -6779,7 +6781,7 @@ _NEON2SSE_INLINE float32x2_t vrecpe_f32(float32x2_t a) //use low 64 bits
     return res64;
 }
 
-uint32x2_t vrecpe_u32(uint32x2_t a); // VRECPE.U32 d0,d0
+_NEON2SSESTORAGE uint32x2_t vrecpe_u32(uint32x2_t a); // VRECPE.U32 d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vrecpe_u32(uint32x2_t a), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //Input is  fixed point number!!! No reciprocal for ints in IA32 available
@@ -6801,21 +6803,21 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vrecpe_u32(uint32x2_t
     return res;
 }
 
-float32x4_t vrecpeq_f32(float32x4_t a); // VRECPE.F32 q0,q0
+_NEON2SSESTORAGE float32x4_t vrecpeq_f32(float32x4_t a); // VRECPE.F32 q0,q0
 #define vrecpeq_f32 _mm_rcp_ps
 
 
-uint32x4_t vrecpeq_u32(uint32x4_t a); // VRECPE.U32 q0,q0
+_NEON2SSESTORAGE uint32x4_t vrecpeq_u32(uint32x4_t a); // VRECPE.U32 q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vrecpeq_u32(uint32x4_t a), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //Input is  fixed point number!!!
     //We implement the recip_estimate function as described in ARMv7 reference manual (VRECPE instruction) but use float instead of double
     _NEON2SSE_ALIGN_16 uint32_t atmp[4];
     _NEON2SSE_ALIGN_16 uint32_t res[4];
-   _NEON2SSE_ALIGN_16 uint32_t c80000000[4] = {0x80000000,0x80000000, 0x80000000,0x80000000};
+    _NEON2SSE_ALIGN_16 static const uint32_t c80000000[4] = {0x80000000,0x80000000, 0x80000000,0x80000000};
     float resf, r;
     int i, q, s;
-  __m128i res128, mask, zero;
+    __m128i res128, mask, zero;
     _mm_store_si128((__m128i*)atmp, a);
     zero = _mm_setzero_si128();
     for (i =0; i<4; i++){
@@ -6837,7 +6839,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vrecpeq_u32(uint32x4_t
 //no reciprocal square root for ints in IA32 available, neither for unsigned int to float4 lanes conversion, so a serial solution looks faster
 //but the particular implementation for vrsqrte_u32 may vary for various ARM compilers
 ////the ARM NEON and x86 SIMD results may be slightly different
-float32x2_t vrsqrte_f32(float32x2_t a); // VRSQRTE.F32 d0,d0
+_NEON2SSESTORAGE float32x2_t vrsqrte_f32(float32x2_t a); // VRSQRTE.F32 d0,d0
 _NEON2SSE_INLINE float32x2_t vrsqrte_f32(float32x2_t a) //use low 64 bits
 {
     float32x4_t res;
@@ -6847,14 +6849,14 @@ _NEON2SSE_INLINE float32x2_t vrsqrte_f32(float32x2_t a) //use low 64 bits
     return res64;
 }
 
-uint32x2_t vrsqrte_u32(uint32x2_t a); // VRSQRTE.U32 d0,d0
-_NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vrsqrte_u32(uint32x2_t a), _NEON2SSE_REASON_SLOW_SERIAL) 
+_NEON2SSESTORAGE uint32x2_t vrsqrte_u32(uint32x2_t a); // VRSQRTE.U32 d0,d0
+_NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vrsqrte_u32(uint32x2_t a), _NEON2SSE_REASON_SLOW_SERIAL)
 {
-    //Input is  fixed point number!!! 
+    //Input is  fixed point number!!!
     //We implement the recip_sqrt_estimate function as described in ARMv7 reference manual (VRSQRTE instruction) but use float instead of double
    uint32x2_t res;
    __m128 tmp;
-    float r, resf, coeff;	
+    float r, resf, coeff;
     int i,q0, s;
     for (i =0; i<2; i++){
         if((a.m64_u32[i] & 0xc0000000) == 0) { //a <=0x3fffffff
@@ -6874,19 +6876,19 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vrsqrte_u32(uint32x2_t
     return res;
 }
 
-float32x4_t vrsqrteq_f32(float32x4_t a); // VRSQRTE.F32 q0,q0
+_NEON2SSESTORAGE float32x4_t vrsqrteq_f32(float32x4_t a); // VRSQRTE.F32 q0,q0
 #define vrsqrteq_f32 _mm_rsqrt_ps
 
-uint32x4_t vrsqrteq_u32(uint32x4_t a); // VRSQRTE.U32 q0,q0
-_NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vrsqrteq_u32(uint32x4_t a), _NEON2SSE_REASON_SLOW_SERIAL) 
+_NEON2SSESTORAGE uint32x4_t vrsqrteq_u32(uint32x4_t a); // VRSQRTE.U32 q0,q0
+_NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vrsqrteq_u32(uint32x4_t a), _NEON2SSE_REASON_SLOW_SERIAL)
 {
-    //Input is  fixed point number!!! 
+    //Input is  fixed point number!!!
     //We implement the recip_sqrt_estimate function as described in ARMv7 reference manual (VRSQRTE instruction) but use float instead of double
    _NEON2SSE_ALIGN_16 uint32_t  atmp[4], res[4];
-   _NEON2SSE_ALIGN_16 uint32_t c_c0000000[4] = {0xc0000000,0xc0000000, 0xc0000000,0xc0000000};
-  __m128 tmp;
-  __m128i res128, mask, zero;
-    float r, resf, coeff;	
+   _NEON2SSE_ALIGN_16 static const uint32_t c_c0000000[4] = {0xc0000000,0xc0000000, 0xc0000000,0xc0000000};
+   __m128 tmp;
+   __m128i res128, mask, zero;
+    float r, resf, coeff;
     int i,q0, s;
     _mm_store_si128((__m128i*)atmp, a);
     zero = _mm_setzero_si128();
@@ -6912,7 +6914,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vrsqrteq_u32(uint32x4_
 //multiplies the elements of one vector by the corresponding elements of another vector,
 //subtracts each of the results from 2, and places the final results into the elements of the destination vector.
 
-float32x2_t vrecps_f32(float32x2_t a, float32x2_t b); // VRECPS.F32 d0, d0, d0
+_NEON2SSESTORAGE float32x2_t vrecps_f32(float32x2_t a, float32x2_t b); // VRECPS.F32 d0, d0, d0
 _NEON2SSE_INLINE float32x2_t vrecps_f32(float32x2_t a, float32x2_t b)
 {
     float32x4_t res;
@@ -6922,7 +6924,7 @@ _NEON2SSE_INLINE float32x2_t vrecps_f32(float32x2_t a, float32x2_t b)
     return res64;
 }
 
-float32x4_t vrecpsq_f32(float32x4_t a, float32x4_t b); // VRECPS.F32 q0, q0, q0
+_NEON2SSESTORAGE float32x4_t vrecpsq_f32(float32x4_t a, float32x4_t b); // VRECPS.F32 q0, q0, q0
 _NEON2SSE_INLINE float32x4_t vrecpsq_f32(float32x4_t a, float32x4_t b) // VRECPS.F32 q0, q0, q0
 {
     __m128 f2, mul;
@@ -6935,7 +6937,7 @@ _NEON2SSE_INLINE float32x4_t vrecpsq_f32(float32x4_t a, float32x4_t b) // VRECPS
 //multiplies the elements of one vector by the corresponding elements of another vector,
 //subtracts each of the results from 3, divides these results by two, and places the final results into the elements of the destination vector.
 
-float32x2_t vrsqrts_f32(float32x2_t a, float32x2_t b); // VRSQRTS.F32 d0, d0, d0
+_NEON2SSESTORAGE float32x2_t vrsqrts_f32(float32x2_t a, float32x2_t b); // VRSQRTS.F32 d0, d0, d0
 _NEON2SSE_INLINE float32x2_t vrsqrts_f32(float32x2_t a, float32x2_t b)
 {
     float32x2_t res;
@@ -6944,7 +6946,7 @@ _NEON2SSE_INLINE float32x2_t vrsqrts_f32(float32x2_t a, float32x2_t b)
     return res;
 }
 
-float32x4_t vrsqrtsq_f32(float32x4_t a, float32x4_t b); // VRSQRTS.F32 q0, q0, q0
+_NEON2SSESTORAGE float32x4_t vrsqrtsq_f32(float32x4_t a, float32x4_t b); // VRSQRTS.F32 q0, q0, q0
 _NEON2SSE_INLINE float32x4_t vrsqrtsq_f32(float32x4_t a, float32x4_t b) // VRSQRTS.F32 q0, q0, q0
 {
     __m128 f3, f05, mul;
@@ -6976,97 +6978,97 @@ _NEON2SSE_INLINE float32x4_t vrsqrtsq_f32(float32x4_t a, float32x4_t b) // VRSQR
         else res.m64_ ## SIGN ## TYPE[i] = (b.m64_i ## TYPE[i] >=0) ? a.m64_ ## SIGN ## TYPE[i] << b.m64_i ## TYPE[i] : a.m64_ ## SIGN ## TYPE[i] >> (-b.m64_i ## TYPE[i]); } \
         return res;
 
-int8x8_t vshl_s8(int8x8_t a, int8x8_t b); // VSHL.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vshl_s8(int8x8_t a, int8x8_t b); // VSHL.S8 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x8_t vshl_s8(int8x8_t a, int8x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT_64(8, i, 8)
 }
 
-int16x4_t vshl_s16(int16x4_t a, int16x4_t b); // VSHL.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vshl_s16(int16x4_t a, int16x4_t b); // VSHL.S16 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x4_t vshl_s16(int16x4_t a, int16x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT_64(16, i, 4)
 }
 
-int32x2_t vshl_s32(int32x2_t a, int32x2_t b); // VSHL.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vshl_s32(int32x2_t a, int32x2_t b); // VSHL.S32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vshl_s32(int32x2_t a, int32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT_64(32, i, 2)
 }
 
-int64x1_t vshl_s64(int64x1_t a, int64x1_t b); // VSHL.S64 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vshl_s64(int64x1_t a, int64x1_t b); // VSHL.S64 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vshl_s64(int64x1_t a, int64x1_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT_64(64, i, 1)
 }
 
-uint8x8_t vshl_u8(uint8x8_t a, int8x8_t b); // VSHL.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vshl_u8(uint8x8_t a, int8x8_t b); // VSHL.U8 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x8_t vshl_u8(uint8x8_t a, int8x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT_64(8, u, 8)
 }
 
-uint16x4_t vshl_u16(uint16x4_t a, int16x4_t b); // VSHL.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vshl_u16(uint16x4_t a, int16x4_t b); // VSHL.s16 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x4_t vshl_u16(uint16x4_t a, int16x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT_64(16, u, 4)
 }
 
-uint32x2_t vshl_u32(uint32x2_t a, int32x2_t b); // VSHL.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vshl_u32(uint32x2_t a, int32x2_t b); // VSHL.U32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vshl_u32(uint32x2_t a, int32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT_64(32, u, 2)
 }
 
-uint64x1_t vshl_u64(uint64x1_t a, int64x1_t b); // VSHL.U64 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vshl_u64(uint64x1_t a, int64x1_t b); // VSHL.U64 d0,d0,d0
 _NEON2SSE_INLINE uint64x1_t vshl_u64(uint64x1_t a, int64x1_t b) //if we use the SERIAL_SHIFT macro need to have the special processing  for large numbers
 {
     SERIAL_SHIFT_64(64, u, 1)
 }
 
-int8x16_t vshlq_s8(int8x16_t a, int8x16_t b); // VSHL.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vshlq_s8(int8x16_t a, int8x16_t b); // VSHL.S8 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x16_t vshlq_s8(int8x16_t a, int8x16_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT(int8_t, int8_t, 16, 16)
 }
 
-int16x8_t vshlq_s16(int16x8_t a, int16x8_t b); // VSHL.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vshlq_s16(int16x8_t a, int16x8_t b); // VSHL.S16 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x8_t vshlq_s16(int16x8_t a, int16x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT(int16_t, int16_t, 8, 8)
 }
 
-int32x4_t vshlq_s32(int32x4_t a, int32x4_t b); // VSHL.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vshlq_s32(int32x4_t a, int32x4_t b); // VSHL.S32 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vshlq_s32(int32x4_t a, int32x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT(int32_t, int32_t, 4, 4)
 }
 
-int64x2_t vshlq_s64(int64x2_t a, int64x2_t b); // VSHL.S64 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vshlq_s64(int64x2_t a, int64x2_t b); // VSHL.S64 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vshlq_s64(int64x2_t a, int64x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT(int64_t, int64_t, 2, 2)
 }
 
-uint8x16_t vshlq_u8(uint8x16_t a, int8x16_t b); // VSHL.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vshlq_u8(uint8x16_t a, int8x16_t b); // VSHL.U8 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x16_t vshlq_u8(uint8x16_t a, int8x16_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT(uint8_t, int8_t, 16, 16)
 }
 
-uint16x8_t vshlq_u16(uint16x8_t a, int16x8_t b); // VSHL.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vshlq_u16(uint16x8_t a, int16x8_t b); // VSHL.s16 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x8_t vshlq_u16(uint16x8_t a, int16x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT(uint16_t, int16_t, 8, 8)
 }
 
-uint32x4_t vshlq_u32(uint32x4_t a, int32x4_t b); // VSHL.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vshlq_u32(uint32x4_t a, int32x4_t b); // VSHL.U32 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vshlq_u32(uint32x4_t a, int32x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT(uint32_t, int32_t, 4, 4)
 }
 
-uint64x2_t vshlq_u64(uint64x2_t a, int64x2_t b); // VSHL.U64 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vshlq_u64(uint64x2_t a, int64x2_t b); // VSHL.U64 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING( uint64x2_t vshlq_u64(uint64x2_t a, int64x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SHIFT(uint64_t, int64_t, 2, 2)
@@ -7140,97 +7142,97 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING( uint64x2_t vshlq_u64(uint64x2_t
                     res.m64_u ## TYPE[i] = ( a.m64_u ## TYPE[i] >= limit) ? res.m64_u ## TYPE[i] = ~((int ## TYPE ## _t) 0) : a.m64_u ## TYPE[i] << b.m64_u ## TYPE[i]; }}}} \
         return res;
 
-int8x8_t vqshl_s8(int8x8_t a, int8x8_t b); // VQSHL.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vqshl_s8(int8x8_t a, int8x8_t b); // VQSHL.S8 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x8_t vqshl_s8(int8x8_t a, int8x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_SIGNED_64(8,8)
 }
 
-int16x4_t vqshl_s16(int16x4_t a, int16x4_t b); // VQSHL.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vqshl_s16(int16x4_t a, int16x4_t b); // VQSHL.S16 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x4_t vqshl_s16(int16x4_t a, int16x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_SIGNED_64(16,4)
 }
 
-int32x2_t vqshl_s32(int32x2_t a, int32x2_t b); // VQSHL.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqshl_s32(int32x2_t a, int32x2_t b); // VQSHL.S32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqshl_s32(int32x2_t a, int32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_SIGNED_64(32,2)
 }
 
-int64x1_t vqshl_s64(int64x1_t a, int64x1_t b); // VQSHL.S64 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vqshl_s64(int64x1_t a, int64x1_t b); // VQSHL.S64 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vqshl_s64(int64x1_t a, int64x1_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_SIGNED_64(64,1)
 }
 
-uint8x8_t vqshl_u8(uint8x8_t a, int8x8_t b); // VQSHL.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vqshl_u8(uint8x8_t a, int8x8_t b); // VQSHL.U8 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x8_t vqshl_u8(uint8x8_t a, int8x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_UNSIGNED_64(8,8)
 }
 
-uint16x4_t vqshl_u16(uint16x4_t a, int16x4_t b); // VQSHL.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vqshl_u16(uint16x4_t a, int16x4_t b); // VQSHL.s16 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x4_t vqshl_u16(uint16x4_t a, int16x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_UNSIGNED_64(16,4)
 }
 
-uint32x2_t vqshl_u32(uint32x2_t a, int32x2_t b); // VQSHL.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vqshl_u32(uint32x2_t a, int32x2_t b); // VQSHL.U32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vqshl_u32(uint32x2_t a, int32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_UNSIGNED_64(32,2)
 }
 
-uint64x1_t vqshl_u64(uint64x1_t a, int64x1_t b); // VQSHL.U64 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vqshl_u64(uint64x1_t a, int64x1_t b); // VQSHL.U64 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vqshl_u64(uint64x1_t a, int64x1_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_UNSIGNED_64(64,1)
 }
 
-int8x16_t vqshlq_s8(int8x16_t a, int8x16_t b); // VQSHL.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vqshlq_s8(int8x16_t a, int8x16_t b); // VQSHL.S8 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x16_t vqshlq_s8(int8x16_t a, int8x16_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_SIGNED(int8_t, 16, 16)
 }
 
-int16x8_t vqshlq_s16(int16x8_t a, int16x8_t b); // VQSHL.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vqshlq_s16(int16x8_t a, int16x8_t b); // VQSHL.S16 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x8_t vqshlq_s16(int16x8_t a, int16x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_SIGNED(int16_t, 8, 8)
 }
 
-int32x4_t vqshlq_s32(int32x4_t a, int32x4_t b); // VQSHL.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vqshlq_s32(int32x4_t a, int32x4_t b); // VQSHL.S32 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqshlq_s32(int32x4_t a, int32x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_SIGNED(int32_t, 4, 4)
 }
 
-int64x2_t vqshlq_s64(int64x2_t a, int64x2_t b); // VQSHL.S64 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vqshlq_s64(int64x2_t a, int64x2_t b); // VQSHL.S64 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqshlq_s64(int64x2_t a, int64x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_SIGNED(int64_t, 2, 2)
 }
 
-uint8x16_t vqshlq_u8(uint8x16_t a, int8x16_t b); // VQSHL.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vqshlq_u8(uint8x16_t a, int8x16_t b); // VQSHL.U8 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x16_t vqshlq_u8(uint8x16_t a, int8x16_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_UNSIGNED(int8_t, 16, 16)
 }
 
-uint16x8_t vqshlq_u16(uint16x8_t a, int16x8_t b); // VQSHL.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vqshlq_u16(uint16x8_t a, int16x8_t b); // VQSHL.s16 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x8_t vqshlq_u16(uint16x8_t a, int16x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_UNSIGNED(int16_t, 8, 8)
 }
 
-uint32x4_t vqshlq_u32(uint32x4_t a, int32x4_t b); // VQSHL.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vqshlq_u32(uint32x4_t a, int32x4_t b); // VQSHL.U32 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vqshlq_u32(uint32x4_t a, int32x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_UNSIGNED(int32_t, 4, 4)
 }
 
-uint64x2_t vqshlq_u64(uint64x2_t a, int64x2_t b); // VQSHL.U64 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vqshlq_u64(uint64x2_t a, int64x2_t b); // VQSHL.U64 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vqshlq_u64(uint64x2_t a, int64x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_SHIFT_UNSIGNED(int64_t, 2, 2)
@@ -7268,97 +7270,97 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vqshlq_u64(uint64x2_t
         return res;
 
 
-int8x8_t vrshl_s8(int8x8_t a, int8x8_t b); // VRSHL.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vrshl_s8(int8x8_t a, int8x8_t b); // VRSHL.S8 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x8_t vrshl_s8(int8x8_t a, int8x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT_64(8,i,8)
 }
 
-int16x4_t vrshl_s16(int16x4_t a, int16x4_t b); // VRSHL.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vrshl_s16(int16x4_t a, int16x4_t b); // VRSHL.S16 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x4_t vrshl_s16(int16x4_t a, int16x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT_64(16,i,4)
 }
 
-int32x2_t vrshl_s32(int32x2_t a, int32x2_t b); // VRSHL.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vrshl_s32(int32x2_t a, int32x2_t b); // VRSHL.S32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vrshl_s32(int32x2_t a, int32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT_64(32,i,2)
 }
 
-int64x1_t vrshl_s64(int64x1_t a, int64x1_t b); // VRSHL.S64 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vrshl_s64(int64x1_t a, int64x1_t b); // VRSHL.S64 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vrshl_s64(int64x1_t a, int64x1_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT_64(64,i,1)
 }
 
-uint8x8_t vrshl_u8(uint8x8_t a, int8x8_t b); // VRSHL.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vrshl_u8(uint8x8_t a, int8x8_t b); // VRSHL.U8 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x8_t vrshl_u8(uint8x8_t a, int8x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT_64(8,u,8)
 }
 
-uint16x4_t vrshl_u16(uint16x4_t a, int16x4_t b); // VRSHL.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vrshl_u16(uint16x4_t a, int16x4_t b); // VRSHL.s16 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x4_t vrshl_u16(uint16x4_t a, int16x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT_64(16,u,4)
 }
 
-uint32x2_t vrshl_u32(uint32x2_t a, int32x2_t b); // VRSHL.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vrshl_u32(uint32x2_t a, int32x2_t b); // VRSHL.U32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vrshl_u32(uint32x2_t a, int32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT_64(32,u,2)
 }
 
-uint64x1_t vrshl_u64(uint64x1_t a, int64x1_t b); // VRSHL.U64 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vrshl_u64(uint64x1_t a, int64x1_t b); // VRSHL.U64 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vrshl_u64(uint64x1_t a, int64x1_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT_64(64,u,1)
 }
 
-int8x16_t vrshlq_s8(int8x16_t a, int8x16_t b); // VRSHL.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vrshlq_s8(int8x16_t a, int8x16_t b); // VRSHL.S8 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x16_t vrshlq_s8(int8x16_t a, int8x16_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT(int8_t, int8_t, 16, 16)
 }
 
-int16x8_t vrshlq_s16(int16x8_t a, int16x8_t b); // VRSHL.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vrshlq_s16(int16x8_t a, int16x8_t b); // VRSHL.S16 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x8_t vrshlq_s16(int16x8_t a, int16x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT(int16_t, int16_t, 8, 8)
 }
 
-int32x4_t vrshlq_s32(int32x4_t a, int32x4_t b); // VRSHL.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vrshlq_s32(int32x4_t a, int32x4_t b); // VRSHL.S32 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vrshlq_s32(int32x4_t a, int32x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT(int32_t, int32_t, 4, 4)
 }
 
-int64x2_t vrshlq_s64(int64x2_t a, int64x2_t b); // VRSHL.S64 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vrshlq_s64(int64x2_t a, int64x2_t b); // VRSHL.S64 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vrshlq_s64(int64x2_t a, int64x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT(int64_t, int64_t, 2, 2)
 }
 
-uint8x16_t vrshlq_u8(uint8x16_t a, int8x16_t b); // VRSHL.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vrshlq_u8(uint8x16_t a, int8x16_t b); // VRSHL.U8 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x16_t vrshlq_u8(uint8x16_t a, int8x16_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT(uint8_t, int8_t, 16, 16)
 }
 
-uint16x8_t vrshlq_u16(uint16x8_t a, int16x8_t b); // VRSHL.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vrshlq_u16(uint16x8_t a, int16x8_t b); // VRSHL.s16 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x8_t vrshlq_u16(uint16x8_t a, int16x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT(uint16_t, int16_t, 8, 8)
 }
 
-uint32x4_t vrshlq_u32(uint32x4_t a, int32x4_t b); // VRSHL.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vrshlq_u32(uint32x4_t a, int32x4_t b); // VRSHL.U32 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vrshlq_u32(uint32x4_t a, int32x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT(uint32_t, int32_t, 4, 4)
 }
 
-uint64x2_t vrshlq_u64(uint64x2_t a, int64x2_t b); // VRSHL.U64 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vrshlq_u64(uint64x2_t a, int64x2_t b); // VRSHL.U64 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vrshlq_u64(uint64x2_t a, int64x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_ROUNDING_SHIFT(uint64_t, int64_t, 2, 2)
@@ -7433,97 +7435,97 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vrshlq_u64(uint64x2_t
                     res.m64_u ## TYPE[i] = ( a.m64_u ## TYPE[i] >= limit) ? res.m64_u ## TYPE[i] = ~((int ## TYPE ## _t) 0) : a.m64_u ## TYPE[i] << b.m64_i ## TYPE[i]; }}}} \
         return res;
 
-int8x8_t vqrshl_s8(int8x8_t a, int8x8_t b); // VQRSHL.S8 d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vqrshl_s8(int8x8_t a, int8x8_t b); // VQRSHL.S8 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x8_t vqrshl_s8(int8x8_t a, int8x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_SIGNED_64(8,8)
 }
 
-int16x4_t vqrshl_s16(int16x4_t a, int16x4_t b); // VQRSHL.S16 d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vqrshl_s16(int16x4_t a, int16x4_t b); // VQRSHL.S16 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x4_t vqrshl_s16(int16x4_t a, int16x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_SIGNED_64(16,4)
 }
 
-int32x2_t vqrshl_s32(int32x2_t a, int32x2_t b); // VQRSHL.S32 d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vqrshl_s32(int32x2_t a, int32x2_t b); // VQRSHL.S32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqrshl_s32(int32x2_t a, int32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_SIGNED_64(32,2)
 }
 
-int64x1_t vqrshl_s64(int64x1_t a, int64x1_t b); // VQRSHL.S64 d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vqrshl_s64(int64x1_t a, int64x1_t b); // VQRSHL.S64 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vqrshl_s64(int64x1_t a, int64x1_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_SIGNED_64(64,1)
 }
 
-uint8x8_t vqrshl_u8(uint8x8_t a, int8x8_t b); // VQRSHL.U8 d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vqrshl_u8(uint8x8_t a, int8x8_t b); // VQRSHL.U8 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x8_t vqrshl_u8(uint8x8_t a, int8x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_UNSIGNED_64(8,8)
 }
 
-uint16x4_t vqrshl_u16(uint16x4_t a, int16x4_t b); // VQRSHL.s16 d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vqrshl_u16(uint16x4_t a, int16x4_t b); // VQRSHL.s16 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x4_t vqrshl_u16(uint16x4_t a, int16x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_UNSIGNED_64(16,4)
 }
 
-uint32x2_t vqrshl_u32(uint32x2_t a, int32x2_t b); // VQRSHL.U32 d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vqrshl_u32(uint32x2_t a, int32x2_t b); // VQRSHL.U32 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vqrshl_u32(uint32x2_t a, int32x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_UNSIGNED_64(32,2)
 }
 
-uint64x1_t vqrshl_u64(uint64x1_t a, int64x1_t b); // VQRSHL.U64 d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vqrshl_u64(uint64x1_t a, int64x1_t b); // VQRSHL.U64 d0,d0,d0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vqrshl_u64(uint64x1_t a, int64x1_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_UNSIGNED_64(64,1)
 }
 
-int8x16_t vqrshlq_s8(int8x16_t a, int8x16_t b); // VQRSHL.S8 q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vqrshlq_s8(int8x16_t a, int8x16_t b); // VQRSHL.S8 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x16_t vqrshlq_s8(int8x16_t a, int8x16_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_SIGNED(int8_t, 16, 16)
 }
 
-int16x8_t vqrshlq_s16(int16x8_t a, int16x8_t b); // VQRSHL.S16 q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vqrshlq_s16(int16x8_t a, int16x8_t b); // VQRSHL.S16 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x8_t vqrshlq_s16(int16x8_t a, int16x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_SIGNED(int16_t, 8, 8)
 }
 
-int32x4_t vqrshlq_s32(int32x4_t a, int32x4_t b); // VQRSHL.S32 q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vqrshlq_s32(int32x4_t a, int32x4_t b); // VQRSHL.S32 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqrshlq_s32(int32x4_t a, int32x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_SIGNED(int32_t, 4, 4)
 }
 
-int64x2_t vqrshlq_s64(int64x2_t a, int64x2_t b); // VQRSHL.S64 q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vqrshlq_s64(int64x2_t a, int64x2_t b); // VQRSHL.S64 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqrshlq_s64(int64x2_t a, int64x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_SIGNED(int64_t, 2, 2)
 }
 
-uint8x16_t vqrshlq_u8(uint8x16_t a, int8x16_t b); // VQRSHL.U8 q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vqrshlq_u8(uint8x16_t a, int8x16_t b); // VQRSHL.U8 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x16_t vqrshlq_u8(uint8x16_t a, int8x16_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_UNSIGNED(int8_t, 16, 16)
 }
 
-uint16x8_t vqrshlq_u16(uint16x8_t a, int16x8_t b); // VQRSHL.s16 q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vqrshlq_u16(uint16x8_t a, int16x8_t b); // VQRSHL.s16 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x8_t vqrshlq_u16(uint16x8_t a, int16x8_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_UNSIGNED(int16_t, 8, 8)
 }
 
-uint32x4_t vqrshlq_u32(uint32x4_t a, int32x4_t b); // VQRSHL.U32 q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vqrshlq_u32(uint32x4_t a, int32x4_t b); // VQRSHL.U32 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x4_t vqrshlq_u32(uint32x4_t a, int32x4_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_UNSIGNED(int32_t, 4, 4)
 }
 
-uint64x2_t vqrshlq_u64(uint64x2_t a, int64x2_t b); // VQRSHL.U64 q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vqrshlq_u64(uint64x2_t a, int64x2_t b); // VQRSHL.U64 q0,q0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vqrshlq_u64(uint64x2_t a, int64x2_t b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     SERIAL_SATURATING_ROUNDING_SHIFT_UNSIGNED(int64_t, 2, 2)
@@ -7534,7 +7536,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vqrshlq_u64(uint64x2_t
 // *********************************************************************************
 //**************** Vector shift right by constant*************************************
 //************************************************************************************
-int8x8_t vshr_n_s8(int8x8_t a, __constrange(1,8) int b); // VSHR.S8 d0,d0,#8
+_NEON2SSESTORAGE int8x8_t vshr_n_s8(int8x8_t a, __constrange(1,8) int b); // VSHR.S8 d0,d0,#8
 _NEON2SSE_INLINE int8x8_t vshr_n_s8(int8x8_t a, __constrange(1,8) int b) // VSHR.S8 d0,d0,#8
 {
     //no 8 bit shift available, go to 16 bit
@@ -7546,7 +7548,7 @@ _NEON2SSE_INLINE int8x8_t vshr_n_s8(int8x8_t a, __constrange(1,8) int b) // VSHR
     return64(r);
 }
 
-int16x4_t vshr_n_s16(int16x4_t a,  __constrange(1,16) int b); // VSHR.S16 d0,d0,#16
+_NEON2SSESTORAGE int16x4_t vshr_n_s16(int16x4_t a,  __constrange(1,16) int b); // VSHR.S16 d0,d0,#16
 _NEON2SSE_INLINE int16x4_t vshr_n_s16(int16x4_t a,  __constrange(1,16) int b)
 {
     int16x4_t res64;
@@ -7554,14 +7556,14 @@ _NEON2SSE_INLINE int16x4_t vshr_n_s16(int16x4_t a,  __constrange(1,16) int b)
 }
 
 
-int32x2_t vshr_n_s32(int32x2_t a,  __constrange(1,32) int b); // VSHR.S32 d0,d0,#32
+_NEON2SSESTORAGE int32x2_t vshr_n_s32(int32x2_t a,  __constrange(1,32) int b); // VSHR.S32 d0,d0,#32
 _NEON2SSE_INLINE int32x2_t vshr_n_s32(int32x2_t a,  __constrange(1,32) int b)
 {
     int32x2_t res64;
     return64(_mm_srai_epi32(_pM128i(a), b));
 }
 
-int64x1_t vshr_n_s64(int64x1_t a, __constrange(1,64) int b); // VSHR.S64 d0,d0,#64
+_NEON2SSESTORAGE int64x1_t vshr_n_s64(int64x1_t a, __constrange(1,64) int b); // VSHR.S64 d0,d0,#64
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vshr_n_s64(int64x1_t a, __constrange(1,64) int b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //no arithmetic shift for 64bit values, serial solution used
@@ -7571,7 +7573,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vshr_n_s64(int64x1_t a,
     return res;
 }
 
-uint8x8_t vshr_n_u8(uint8x8_t a, __constrange(1,8) int b); // VSHR.U8 d0,d0,#8
+_NEON2SSESTORAGE uint8x8_t vshr_n_u8(uint8x8_t a, __constrange(1,8) int b); // VSHR.U8 d0,d0,#8
 _NEON2SSE_INLINE uint8x8_t vshr_n_u8(uint8x8_t a, __constrange(1,8) int b) // VSHR.U8 d0,d0,#8
 {
     //no 8 bit shift available, go to 16 bit
@@ -7583,7 +7585,7 @@ _NEON2SSE_INLINE uint8x8_t vshr_n_u8(uint8x8_t a, __constrange(1,8) int b) // VS
     return64(r);
 }
 
-uint16x4_t vshr_n_u16(uint16x4_t a,  __constrange(1,16) int b); // VSHR.s16 d0,d0,#16
+_NEON2SSESTORAGE uint16x4_t vshr_n_u16(uint16x4_t a,  __constrange(1,16) int b); // VSHR.s16 d0,d0,#16
 _NEON2SSE_INLINE uint16x4_t vshr_n_u16(uint16x4_t a,  __constrange(1,16) int b)
 {
     uint16x4_t res64;
@@ -7591,7 +7593,7 @@ _NEON2SSE_INLINE uint16x4_t vshr_n_u16(uint16x4_t a,  __constrange(1,16) int b)
 }
 
 
-uint32x2_t vshr_n_u32(uint32x2_t a,  __constrange(1,32) int b); // VSHR.U32 d0,d0,#32
+_NEON2SSESTORAGE uint32x2_t vshr_n_u32(uint32x2_t a,  __constrange(1,32) int b); // VSHR.U32 d0,d0,#32
 _NEON2SSE_INLINE uint32x2_t vshr_n_u32(uint32x2_t a,  __constrange(1,32) int b)
 {
     uint32x2_t res64;
@@ -7599,7 +7601,7 @@ _NEON2SSE_INLINE uint32x2_t vshr_n_u32(uint32x2_t a,  __constrange(1,32) int b)
 }
 
 
-uint64x1_t vshr_n_u64(uint64x1_t a,  __constrange(1,64) int b); // VSHR.U64 d0,d0,#64
+_NEON2SSESTORAGE uint64x1_t vshr_n_u64(uint64x1_t a,  __constrange(1,64) int b); // VSHR.U64 d0,d0,#64
 _NEON2SSE_INLINE uint64x1_t vshr_n_u64(uint64x1_t a,  __constrange(1,64) int b)
 {
     uint64x1_t res64;
@@ -7607,12 +7609,12 @@ _NEON2SSE_INLINE uint64x1_t vshr_n_u64(uint64x1_t a,  __constrange(1,64) int b)
 }
 
 
-int8x16_t vshrq_n_s8(int8x16_t a, __constrange(1,8) int b); // VSHR.S8 q0,q0,#8
+_NEON2SSESTORAGE int8x16_t vshrq_n_s8(int8x16_t a, __constrange(1,8) int b); // VSHR.S8 q0,q0,#8
 _NEON2SSE_INLINE int8x16_t vshrq_n_s8(int8x16_t a, __constrange(1,8) int b) // VSHR.S8 q0,q0,#8
 {
     //no 8 bit shift available, go to 16 bit trick
     __m128i zero, mask0, a_sign, r, a_sign_mask;
-    _NEON2SSE_ALIGN_16 int16_t mask0_16[9] = {0x0000, 0x0080, 0x00c0, 0x00e0, 0x00f0,  0x00f8, 0x00fc, 0x00fe, 0x00ff};
+    _NEON2SSE_ALIGN_16 static const int16_t mask0_16[9] = {0x0000, 0x0080, 0x00c0, 0x00e0, 0x00f0,  0x00f8, 0x00fc, 0x00fe, 0x00ff};
     zero = _mm_setzero_si128();
     mask0 = _mm_set1_epi16(mask0_16[b]); //to mask the bits to be "spoiled"  by 16 bit shift
     a_sign =  _mm_cmpgt_epi8 (zero, a); //ff if a<0 or zero if a>0
@@ -7622,18 +7624,18 @@ _NEON2SSE_INLINE int8x16_t vshrq_n_s8(int8x16_t a, __constrange(1,8) int b) // V
     return _mm_or_si128 (r, a_sign_mask);
 }
 
-int16x8_t vshrq_n_s16(int16x8_t a, __constrange(1,16) int b); // VSHR.S16 q0,q0,#16
+_NEON2SSESTORAGE int16x8_t vshrq_n_s16(int16x8_t a, __constrange(1,16) int b); // VSHR.S16 q0,q0,#16
 #define vshrq_n_s16 _mm_srai_epi16
 
-int32x4_t vshrq_n_s32(int32x4_t a, __constrange(1,32) int b); // VSHR.S32 q0,q0,#32
+_NEON2SSESTORAGE int32x4_t vshrq_n_s32(int32x4_t a, __constrange(1,32) int b); // VSHR.S32 q0,q0,#32
 #define vshrq_n_s32 _mm_srai_epi32
 
-int64x2_t vshrq_n_s64(int64x2_t a, __constrange(1,64) int b); // VSHR.S64 q0,q0,#64
+_NEON2SSESTORAGE int64x2_t vshrq_n_s64(int64x2_t a, __constrange(1,64) int b); // VSHR.S64 q0,q0,#64
 _NEON2SSE_INLINE int64x2_t vshrq_n_s64(int64x2_t a, __constrange(1,64) int b)
 {
     //SIMD implementation may be not optimal due to 64 bit arithmetic shift absense in x86 SIMD
     __m128i c1, signmask,a0,  res64;
-    _NEON2SSE_ALIGN_16 uint64_t mask[] = {0x8000000000000000, 0x8000000000000000};
+    _NEON2SSE_ALIGN_16 static const uint64_t mask[] = {0x8000000000000000, 0x8000000000000000};
     c1 =  _mm_cmpeq_epi32(a,a); //0xffffffffffffffff
     signmask  =  _mm_slli_epi64 (c1, (64 - b));
     a0 = _mm_or_si128(a, *(__m128i*)mask); //get the first bit
@@ -7643,29 +7645,29 @@ _NEON2SSE_INLINE int64x2_t vshrq_n_s64(int64x2_t a, __constrange(1,64) int b)
     return _mm_or_si128(res64, signmask);
 }
 
-uint8x16_t vshrq_n_u8(uint8x16_t a, __constrange(1,8) int b); // VSHR.U8 q0,q0,#8
+_NEON2SSESTORAGE uint8x16_t vshrq_n_u8(uint8x16_t a, __constrange(1,8) int b); // VSHR.U8 q0,q0,#8
 _NEON2SSE_INLINE uint8x16_t vshrq_n_u8(uint8x16_t a, __constrange(1,8) int b) // VSHR.U8 q0,q0,#8
 {
     //no 8 bit shift available, need the special trick
     __m128i mask0, r;
-    _NEON2SSE_ALIGN_16 uint16_t mask10_16[9] = {0xffff, 0xff7f, 0xff3f, 0xff1f, 0xff0f,  0xff07, 0xff03, 0xff01, 0xff00};
+    _NEON2SSE_ALIGN_16 static const uint16_t mask10_16[9] = {0xffff, 0xff7f, 0xff3f, 0xff1f, 0xff0f,  0xff07, 0xff03, 0xff01, 0xff00};
     mask0 = _mm_set1_epi16(mask10_16[b]); //to mask the bits to be "spoiled"  by 16 bit shift
     r = _mm_srli_epi16 ( a, b);
     return _mm_and_si128 (r,  mask0);
 }
 
-uint16x8_t vshrq_n_u16(uint16x8_t a, __constrange(1,16) int b); // VSHR.s16 q0,q0,#16
+_NEON2SSESTORAGE uint16x8_t vshrq_n_u16(uint16x8_t a, __constrange(1,16) int b); // VSHR.s16 q0,q0,#16
 #define vshrq_n_u16 _mm_srli_epi16
 
-uint32x4_t vshrq_n_u32(uint32x4_t a, __constrange(1,32) int b); // VSHR.U32 q0,q0,#32
+_NEON2SSESTORAGE uint32x4_t vshrq_n_u32(uint32x4_t a, __constrange(1,32) int b); // VSHR.U32 q0,q0,#32
 #define vshrq_n_u32 _mm_srli_epi32
 
-uint64x2_t vshrq_n_u64(uint64x2_t a, __constrange(1,64) int b); // VSHR.U64 q0,q0,#64
+_NEON2SSESTORAGE uint64x2_t vshrq_n_u64(uint64x2_t a, __constrange(1,64) int b); // VSHR.U64 q0,q0,#64
 #define vshrq_n_u64 _mm_srli_epi64
 
 //*************************** Vector shift left by constant *************************
 //*********************************************************************************
-int8x8_t vshl_n_s8(int8x8_t a, __constrange(0,7) int b); // VSHL.I8 d0,d0,#0
+_NEON2SSESTORAGE int8x8_t vshl_n_s8(int8x8_t a, __constrange(0,7) int b); // VSHL.I8 d0,d0,#0
 _NEON2SSE_INLINE int8x8_t vshl_n_s8(int8x8_t a, __constrange(0,7) int b) // VSHL.I8 d0,d0,#0
 {
     //no 8 bit shift available, go to 16 bit
@@ -7677,7 +7679,7 @@ _NEON2SSE_INLINE int8x8_t vshl_n_s8(int8x8_t a, __constrange(0,7) int b) // VSHL
     return64(r);
 }
 
-int16x4_t vshl_n_s16(int16x4_t a,  __constrange(0,15) int b); // VSHL.I16 d0,d0,#0
+_NEON2SSESTORAGE int16x4_t vshl_n_s16(int16x4_t a,  __constrange(0,15) int b); // VSHL.I16 d0,d0,#0
 _NEON2SSE_INLINE int16x4_t vshl_n_s16(int16x4_t a,  __constrange(0,15) int b)
 {
     int16x4_t res64;
@@ -7685,7 +7687,7 @@ _NEON2SSE_INLINE int16x4_t vshl_n_s16(int16x4_t a,  __constrange(0,15) int b)
 }
 
 
-int32x2_t vshl_n_s32(int32x2_t a,  __constrange(0,31) int b); // VSHL.I32 d0,d0,#0
+_NEON2SSESTORAGE int32x2_t vshl_n_s32(int32x2_t a,  __constrange(0,31) int b); // VSHL.I32 d0,d0,#0
 _NEON2SSE_INLINE int32x2_t vshl_n_s32(int32x2_t a,  __constrange(0,31) int b)
 {
     int32x2_t res64;
@@ -7693,7 +7695,7 @@ _NEON2SSE_INLINE int32x2_t vshl_n_s32(int32x2_t a,  __constrange(0,31) int b)
 }
 
 
-int64x1_t vshl_n_s64(int64x1_t a,  __constrange(0,63) int b); // VSHL.I64 d0,d0,#0
+_NEON2SSESTORAGE int64x1_t vshl_n_s64(int64x1_t a,  __constrange(0,63) int b); // VSHL.I64 d0,d0,#0
 _NEON2SSE_INLINE int64x1_t vshl_n_s64(int64x1_t a,  __constrange(0,63) int b)
 {
     int64x1_t res64;
@@ -7701,7 +7703,7 @@ _NEON2SSE_INLINE int64x1_t vshl_n_s64(int64x1_t a,  __constrange(0,63) int b)
 }
 
 
-uint8x8_t vshl_n_u8(uint8x8_t a, __constrange(0,7) int b); // VSHL.I8 d0,d0,#0
+_NEON2SSESTORAGE uint8x8_t vshl_n_u8(uint8x8_t a, __constrange(0,7) int b); // VSHL.I8 d0,d0,#0
 _NEON2SSE_INLINE uint8x8_t vshl_n_u8(uint8x8_t a, __constrange(0,7) int b)
 {
     //no 8 bit shift available, go to 16 bit
@@ -7716,52 +7718,52 @@ _NEON2SSE_INLINE uint8x8_t vshl_n_u8(uint8x8_t a, __constrange(0,7) int b)
     return64(r);
 }
 
-uint16x4_t vshl_n_u16(uint16x4_t a,  __constrange(0,15) int b); // VSHL.I16 d0,d0,#0
+_NEON2SSESTORAGE uint16x4_t vshl_n_u16(uint16x4_t a,  __constrange(0,15) int b); // VSHL.I16 d0,d0,#0
 #define vshl_n_u16 vshl_n_s16
 
 
-uint32x2_t vshl_n_u32(uint32x2_t a,  __constrange(0,31) int b); // VSHL.I32 d0,d0,#0
+_NEON2SSESTORAGE uint32x2_t vshl_n_u32(uint32x2_t a,  __constrange(0,31) int b); // VSHL.I32 d0,d0,#0
 #define vshl_n_u32 vshl_n_s32
 
-uint64x1_t vshl_n_u64(uint64x1_t a, __constrange(0,63) int b); // VSHL.I64 d0,d0,#0
+_NEON2SSESTORAGE uint64x1_t vshl_n_u64(uint64x1_t a, __constrange(0,63) int b); // VSHL.I64 d0,d0,#0
 #define vshl_n_u64 vshl_n_s64
 
-int8x16_t vshlq_n_s8(int8x16_t a, __constrange(0,7) int b); // VSHL.I8 q0,q0,#0
+_NEON2SSESTORAGE int8x16_t vshlq_n_s8(int8x16_t a, __constrange(0,7) int b); // VSHL.I8 q0,q0,#0
 #define vshlq_n_s8 vshlq_n_u8
 
-int16x8_t vshlq_n_s16(int16x8_t a, __constrange(0,15) int b); // VSHL.I16 q0,q0,#0
+_NEON2SSESTORAGE int16x8_t vshlq_n_s16(int16x8_t a, __constrange(0,15) int b); // VSHL.I16 q0,q0,#0
 #define vshlq_n_s16 _mm_slli_epi16
 
-int32x4_t vshlq_n_s32(int32x4_t a, __constrange(0,31) int b); // VSHL.I32 q0,q0,#0
+_NEON2SSESTORAGE int32x4_t vshlq_n_s32(int32x4_t a, __constrange(0,31) int b); // VSHL.I32 q0,q0,#0
 #define vshlq_n_s32 _mm_slli_epi32
 
-int64x2_t vshlq_n_s64(int64x2_t a, __constrange(0,63) int b); // VSHL.I64 q0,q0,#0
+_NEON2SSESTORAGE int64x2_t vshlq_n_s64(int64x2_t a, __constrange(0,63) int b); // VSHL.I64 q0,q0,#0
 #define vshlq_n_s64 _mm_slli_epi64
 
-uint8x16_t vshlq_n_u8(uint8x16_t a, __constrange(0,7) int b); // VSHL.I8 q0,q0,#0
+_NEON2SSESTORAGE uint8x16_t vshlq_n_u8(uint8x16_t a, __constrange(0,7) int b); // VSHL.I8 q0,q0,#0
 _NEON2SSE_INLINE uint8x16_t vshlq_n_u8(uint8x16_t a, __constrange(0,7) int b)
 {
     //no 8 bit shift available, need the special trick
     __m128i mask0, r;
-    _NEON2SSE_ALIGN_16 uint16_t mask10_16[9] = {0xffff, 0xfeff, 0xfcff, 0xf8ff, 0xf0ff,  0xe0ff, 0xc0ff, 0x80ff, 0xff};
+    _NEON2SSE_ALIGN_16 static const uint16_t mask10_16[9] = {0xffff, 0xfeff, 0xfcff, 0xf8ff, 0xf0ff,  0xe0ff, 0xc0ff, 0x80ff, 0xff};
     mask0 = _mm_set1_epi16(mask10_16[b]); //to mask the bits to be "spoiled"  by 16 bit shift
     r = _mm_slli_epi16 ( a, b);
     return _mm_and_si128 (r,  mask0);
 }
 
-uint16x8_t vshlq_n_u16(uint16x8_t a, __constrange(0,15) int b); // VSHL.I16 q0,q0,#0
+_NEON2SSESTORAGE uint16x8_t vshlq_n_u16(uint16x8_t a, __constrange(0,15) int b); // VSHL.I16 q0,q0,#0
 #define vshlq_n_u16 vshlq_n_s16
 
-uint32x4_t vshlq_n_u32(uint32x4_t a, __constrange(0,31) int b); // VSHL.I32 q0,q0,#0
+_NEON2SSESTORAGE uint32x4_t vshlq_n_u32(uint32x4_t a, __constrange(0,31) int b); // VSHL.I32 q0,q0,#0
 #define vshlq_n_u32 vshlq_n_s32
 
-uint64x2_t vshlq_n_u64(uint64x2_t a, __constrange(0,63) int b); // VSHL.I64 q0,q0,#0
+_NEON2SSESTORAGE uint64x2_t vshlq_n_u64(uint64x2_t a, __constrange(0,63) int b); // VSHL.I64 q0,q0,#0
 #define vshlq_n_u64 vshlq_n_s64
 
 //************* Vector rounding shift right by constant ******************
 //*************************************************************************
 //No corresponding  x86 intrinsics exist, need to do some tricks
-int8x8_t vrshr_n_s8(int8x8_t a, __constrange(1,8) int b); // VRSHR.S8 d0,d0,#8
+_NEON2SSESTORAGE int8x8_t vrshr_n_s8(int8x8_t a, __constrange(1,8) int b); // VRSHR.S8 d0,d0,#8
 _NEON2SSE_INLINE int8x8_t vrshr_n_s8(int8x8_t a, __constrange(1,8) int b) // VRSHR.S8 d0,d0,#8
 {
     //no 8 bit shift available, go to 16 bit
@@ -7776,7 +7778,7 @@ _NEON2SSE_INLINE int8x8_t vrshr_n_s8(int8x8_t a, __constrange(1,8) int b) // VRS
     return64(r);
 }
 
-int16x4_t vrshr_n_s16(int16x4_t a,  __constrange(1,16) int b); // VRSHR.S16 d0,d0,#16
+_NEON2SSESTORAGE int16x4_t vrshr_n_s16(int16x4_t a,  __constrange(1,16) int b); // VRSHR.S16 d0,d0,#16
 _NEON2SSE_INLINE int16x4_t vrshr_n_s16(int16x4_t a,  __constrange(1,16) int b)
 {
     int16x4_t res64;
@@ -7784,7 +7786,7 @@ _NEON2SSE_INLINE int16x4_t vrshr_n_s16(int16x4_t a,  __constrange(1,16) int b)
 }
 
 
-int32x2_t vrshr_n_s32(int32x2_t a,  __constrange(1,32) int b); // VRSHR.S32 d0,d0,#32
+_NEON2SSESTORAGE int32x2_t vrshr_n_s32(int32x2_t a,  __constrange(1,32) int b); // VRSHR.S32 d0,d0,#32
 _NEON2SSE_INLINE int32x2_t vrshr_n_s32(int32x2_t a,  __constrange(1,32) int b)
 {
     int32x2_t res64;
@@ -7792,7 +7794,7 @@ _NEON2SSE_INLINE int32x2_t vrshr_n_s32(int32x2_t a,  __constrange(1,32) int b)
 }
 
 
-int64x1_t vrshr_n_s64(int64x1_t a, __constrange(1,64) int b); // VRSHR.S64 d0,d0,#64
+_NEON2SSESTORAGE int64x1_t vrshr_n_s64(int64x1_t a, __constrange(1,64) int b); // VRSHR.S64 d0,d0,#64
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vrshr_n_s64(int64x1_t a, __constrange(1,64) int b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     //serial solution is faster
@@ -7807,7 +7809,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vrshr_n_s64(int64x1_t a
     return res;
 }
 
-uint8x8_t vrshr_n_u8(uint8x8_t a, __constrange(1,8) int b); // VRSHR.U8 d0,d0,#8
+_NEON2SSESTORAGE uint8x8_t vrshr_n_u8(uint8x8_t a, __constrange(1,8) int b); // VRSHR.U8 d0,d0,#8
 _NEON2SSE_INLINE uint8x8_t vrshr_n_u8(uint8x8_t a, __constrange(1,8) int b) // VRSHR.U8 d0,d0,#8
 {
     //no 8 bit shift available, go to 16 bit, solution may be not optimal compared with the serial one
@@ -7822,7 +7824,7 @@ _NEON2SSE_INLINE uint8x8_t vrshr_n_u8(uint8x8_t a, __constrange(1,8) int b) // V
     return64(r);
 }
 
-uint16x4_t vrshr_n_u16(uint16x4_t a,  __constrange(1,16) int b); // VRSHR.s16 d0,d0,#16
+_NEON2SSESTORAGE uint16x4_t vrshr_n_u16(uint16x4_t a,  __constrange(1,16) int b); // VRSHR.s16 d0,d0,#16
 _NEON2SSE_INLINE uint16x4_t vrshr_n_u16(uint16x4_t a,  __constrange(1,16) int b)
 {
     uint16x4_t res64;
@@ -7830,7 +7832,7 @@ _NEON2SSE_INLINE uint16x4_t vrshr_n_u16(uint16x4_t a,  __constrange(1,16) int b)
 }
 
 
-uint32x2_t vrshr_n_u32(uint32x2_t a,  __constrange(1,32) int b); // VRSHR.U32 d0,d0,#32
+_NEON2SSESTORAGE uint32x2_t vrshr_n_u32(uint32x2_t a,  __constrange(1,32) int b); // VRSHR.U32 d0,d0,#32
 _NEON2SSE_INLINE uint32x2_t vrshr_n_u32(uint32x2_t a,  __constrange(1,32) int b)
 {
     uint32x2_t res64;
@@ -7838,19 +7840,19 @@ _NEON2SSE_INLINE uint32x2_t vrshr_n_u32(uint32x2_t a,  __constrange(1,32) int b)
 }
 
 
-uint64x1_t vrshr_n_u64(uint64x1_t a, __constrange(1,64) int b); // VRSHR.U64 d0,d0,#64
+_NEON2SSESTORAGE uint64x1_t vrshr_n_u64(uint64x1_t a, __constrange(1,64) int b); // VRSHR.U64 d0,d0,#64
 _NEON2SSE_INLINE uint64x1_t vrshr_n_u64(uint64x1_t a, __constrange(1,64) int b)
 {
     uint64x1_t res64;
     return64(vrshrq_n_u64(_pM128i(a), b));
 }
 
-int8x16_t vrshrq_n_s8(int8x16_t a, __constrange(1,8) int b); // VRSHR.S8 q0,q0,#8
+_NEON2SSESTORAGE int8x16_t vrshrq_n_s8(int8x16_t a, __constrange(1,8) int b); // VRSHR.S8 q0,q0,#8
 _NEON2SSE_INLINE int8x16_t vrshrq_n_s8(int8x16_t a, __constrange(1,8) int b) // VRSHR.S8 q0,q0,#8
 {
     //no 8 bit shift available, go to 16 bit trick
     __m128i r, mask1, maskb;
-    _NEON2SSE_ALIGN_16 uint16_t mask2b[9] = {0x0000, 0x0101, 0x0202, 0x0404, 0x0808, 0x1010, 0x2020, 0x4040, 0x8080}; // 2^b-th bit set to 1
+    _NEON2SSE_ALIGN_16 static const uint16_t mask2b[9] = {0x0000, 0x0101, 0x0202, 0x0404, 0x0808, 0x1010, 0x2020, 0x4040, 0x8080}; // 2^b-th bit set to 1
     r = vshrq_n_s8 (a, b);
     mask1 = _mm_set1_epi16(mask2b[b]); // 2^b-th bit set to 1 for 16bit, need it for rounding
     maskb = _mm_and_si128(a, mask1); //get b or 0 for rounding
@@ -7858,7 +7860,7 @@ _NEON2SSE_INLINE int8x16_t vrshrq_n_s8(int8x16_t a, __constrange(1,8) int b) //
     return _mm_add_epi8(r, maskb); //actual rounding
 }
 
-int16x8_t vrshrq_n_s16(int16x8_t a, __constrange(1,16) int b); // VRSHR.S16 q0,q0,#16
+_NEON2SSESTORAGE int16x8_t vrshrq_n_s16(int16x8_t a, __constrange(1,16) int b); // VRSHR.S16 q0,q0,#16
 _NEON2SSE_INLINE int16x8_t vrshrq_n_s16(int16x8_t a, __constrange(1,16) int b) // VRSHR.S16 q0,q0,#16
 {
     __m128i maskb, r;
@@ -7868,7 +7870,7 @@ _NEON2SSE_INLINE int16x8_t vrshrq_n_s16(int16x8_t a, __constrange(1,16) int b) /
     return _mm_add_epi16 (r, maskb); //actual rounding
 }
 
-int32x4_t vrshrq_n_s32(int32x4_t a, __constrange(1,32) int b); // VRSHR.S32 q0,q0,#32
+_NEON2SSESTORAGE int32x4_t vrshrq_n_s32(int32x4_t a, __constrange(1,32) int b); // VRSHR.S32 q0,q0,#32
 _NEON2SSE_INLINE int32x4_t vrshrq_n_s32(int32x4_t a, __constrange(1,32) int b) // VRSHR.S32 q0,q0,#32
 {
     __m128i maskb,  r;
@@ -7878,7 +7880,7 @@ _NEON2SSE_INLINE int32x4_t vrshrq_n_s32(int32x4_t a, __constrange(1,32) int b) /
     return _mm_add_epi32 (r, maskb); //actual rounding
 }
 
-int64x2_t vrshrq_n_s64(int64x2_t a, __constrange(1,64) int b); // VRSHR.S64 q0,q0,#64
+_NEON2SSESTORAGE int64x2_t vrshrq_n_s64(int64x2_t a, __constrange(1,64) int b); // VRSHR.S64 q0,q0,#64
 _NEON2SSE_INLINE int64x2_t vrshrq_n_s64(int64x2_t a, __constrange(1,64) int b)
 {
     //solution may be not optimal compared with a serial one
@@ -7890,12 +7892,12 @@ _NEON2SSE_INLINE int64x2_t vrshrq_n_s64(int64x2_t a, __constrange(1,64) int b)
     return _mm_add_epi64 (r, maskb); //actual rounding
 }
 
-uint8x16_t vrshrq_n_u8(uint8x16_t a, __constrange(1,8) int b); // VRSHR.U8 q0,q0,#8
+_NEON2SSESTORAGE uint8x16_t vrshrq_n_u8(uint8x16_t a, __constrange(1,8) int b); // VRSHR.U8 q0,q0,#8
 _NEON2SSE_INLINE uint8x16_t vrshrq_n_u8(uint8x16_t a, __constrange(1,8) int b) // VRSHR.U8 q0,q0,#8
 {
     //no 8 bit shift available, go to 16 bit trick
     __m128i r, mask1, maskb;
-    _NEON2SSE_ALIGN_16 uint16_t mask2b[9] = {0x0000, 0x0101, 0x0202, 0x0404, 0x0808, 0x1010, 0x2020, 0x4040, 0x8080}; // 2^b-th bit set to 1
+    _NEON2SSE_ALIGN_16 static const uint16_t mask2b[9] = {0x0000, 0x0101, 0x0202, 0x0404, 0x0808, 0x1010, 0x2020, 0x4040, 0x8080}; // 2^b-th bit set to 1
     r = vshrq_n_u8 (a, b);
     mask1 = _mm_set1_epi16(mask2b[b]); // 2^b-th bit set to 1 for 16bit, need it for rounding
     maskb = _mm_and_si128(a, mask1); //get b or 0 for rounding
@@ -7903,7 +7905,7 @@ _NEON2SSE_INLINE uint8x16_t vrshrq_n_u8(uint8x16_t a, __constrange(1,8) int b) /
     return _mm_add_epi8(r, maskb); //actual rounding
 }
 
-uint16x8_t vrshrq_n_u16(uint16x8_t a, __constrange(1,16) int b); // VRSHR.s16 q0,q0,#16
+_NEON2SSESTORAGE uint16x8_t vrshrq_n_u16(uint16x8_t a, __constrange(1,16) int b); // VRSHR.s16 q0,q0,#16
 _NEON2SSE_INLINE uint16x8_t vrshrq_n_u16(uint16x8_t a, __constrange(1,16) int b) // VRSHR.S16 q0,q0,#16
 {
     __m128i maskb, r;
@@ -7913,7 +7915,7 @@ _NEON2SSE_INLINE uint16x8_t vrshrq_n_u16(uint16x8_t a, __constrange(1,16) int b)
     return _mm_add_epi16 (r, maskb); //actual rounding
 }
 
-uint32x4_t vrshrq_n_u32(uint32x4_t a, __constrange(1,32) int b); // VRSHR.U32 q0,q0,#32
+_NEON2SSESTORAGE uint32x4_t vrshrq_n_u32(uint32x4_t a, __constrange(1,32) int b); // VRSHR.U32 q0,q0,#32
 _NEON2SSE_INLINE uint32x4_t vrshrq_n_u32(uint32x4_t a, __constrange(1,32) int b) // VRSHR.S32 q0,q0,#32
 {
     __m128i maskb,  r;
@@ -7923,7 +7925,7 @@ _NEON2SSE_INLINE uint32x4_t vrshrq_n_u32(uint32x4_t a, __constrange(1,32) int b)
     return _mm_add_epi32 (r, maskb); //actual rounding
 }
 
-uint64x2_t vrshrq_n_u64(uint64x2_t a, __constrange(1,64) int b); // VRSHR.U64 q0,q0,#64
+_NEON2SSESTORAGE uint64x2_t vrshrq_n_u64(uint64x2_t a, __constrange(1,64) int b); // VRSHR.U64 q0,q0,#64
 _NEON2SSE_INLINE uint64x2_t vrshrq_n_u64(uint64x2_t a, __constrange(1,64) int b)
 {
     //solution may be not optimal compared with a serial one
@@ -7936,7 +7938,7 @@ _NEON2SSE_INLINE uint64x2_t vrshrq_n_u64(uint64x2_t a, __constrange(1,64) int b)
 
 //************* Vector shift right by constant and accumulate *********
 //*********************************************************************
-int8x8_t vsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c); // VSRA.S8 d0,d0,#8
+_NEON2SSESTORAGE int8x8_t vsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c); // VSRA.S8 d0,d0,#8
 _NEON2SSE_INLINE int8x8_t vsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c) // VSRA.S8 d0,d0,#8
 {
     int8x8_t shift;
@@ -7944,7 +7946,7 @@ _NEON2SSE_INLINE int8x8_t vsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) in
     return vadd_s8( a, shift);
 }
 
-int16x4_t vsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c); // VSRA.S16 d0,d0,#16
+_NEON2SSESTORAGE int16x4_t vsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c); // VSRA.S16 d0,d0,#16
 _NEON2SSE_INLINE int16x4_t vsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c) // VSRA.S16 d0,d0,#16
 {
     int16x4_t shift;
@@ -7952,7 +7954,7 @@ _NEON2SSE_INLINE int16x4_t vsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,1
     return vadd_s16(a, shift);
 }
 
-int32x2_t vsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c); // VSRA.S32 d0,d0,#32
+_NEON2SSESTORAGE int32x2_t vsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c); // VSRA.S32 d0,d0,#32
 _NEON2SSE_INLINE int32x2_t vsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c) // VSRA.S32 d0,d0,#32
 {
     //may be not optimal compared with the serial execution
@@ -7961,7 +7963,7 @@ _NEON2SSE_INLINE int32x2_t vsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,3
     return vadd_s32( a, shift);
 }
 
-int64x1_t vsra_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VSRA.S64 d0,d0,#64
+_NEON2SSESTORAGE int64x1_t vsra_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VSRA.S64 d0,d0,#64
 _NEON2SSE_INLINE int64x1_t vsra_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c)
 {
     //may be not optimal compared with a serial solution
@@ -7970,7 +7972,7 @@ _NEON2SSE_INLINE int64x1_t vsra_n_s64(int64x1_t a, int64x1_t b, __constrange(1,6
     return vadd_s64( a, shift);
 }
 
-uint8x8_t vsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c); // VSRA.U8 d0,d0,#8
+_NEON2SSESTORAGE uint8x8_t vsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c); // VSRA.U8 d0,d0,#8
 _NEON2SSE_INLINE uint8x8_t vsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c) // VSRA.U8 d0,d0,#8
 {
     uint8x8_t shift;
@@ -7978,7 +7980,7 @@ _NEON2SSE_INLINE uint8x8_t vsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8)
     return vadd_u8(a, shift);
 }
 
-uint16x4_t vsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c); // VSRA.s16 d0,d0,#16
+_NEON2SSESTORAGE uint16x4_t vsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c); // VSRA.s16 d0,d0,#16
 _NEON2SSE_INLINE uint16x4_t vsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c) // VSRA.s16 d0,d0,#16
 {
     uint16x4_t shift;
@@ -7986,7 +7988,7 @@ _NEON2SSE_INLINE uint16x4_t vsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange(
     return vadd_u16(a,shift);
 }
 
-uint32x2_t vsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c); // VSRA.U32 d0,d0,#32
+_NEON2SSESTORAGE uint32x2_t vsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c); // VSRA.U32 d0,d0,#32
 _NEON2SSE_INLINE uint32x2_t vsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c) // VSRA.U32 d0,d0,#32
 {
     //may be not optimal compared with the serial execution
@@ -7995,7 +7997,7 @@ _NEON2SSE_INLINE uint32x2_t vsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange(
     return vadd_u32( a, shift);
 }
 
-uint64x1_t vsra_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VSRA.U64 d0,d0,#64
+_NEON2SSESTORAGE uint64x1_t vsra_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VSRA.U64 d0,d0,#64
 _NEON2SSE_INLINE uint64x1_t vsra_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c) // VSRA.U64 d0,d0,#64
 {
     //may be not optimal compared with the serial execution
@@ -8004,7 +8006,7 @@ _NEON2SSE_INLINE uint64x1_t vsra_n_u64(uint64x1_t a, uint64x1_t b, __constrange(
     return vadd_u64(a, shift);
 }
 
-int8x16_t vsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VSRA.S8 q0,q0,#8
+_NEON2SSESTORAGE int8x16_t vsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VSRA.S8 q0,q0,#8
 _NEON2SSE_INLINE int8x16_t vsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c) // VSRA.S8 q0,q0,#8
 {
     int8x16_t shift;
@@ -8012,7 +8014,7 @@ _NEON2SSE_INLINE int8x16_t vsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8
     return vaddq_s8(a, shift);
 }
 
-int16x8_t vsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VSRA.S16 q0,q0,#16
+_NEON2SSESTORAGE int16x8_t vsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VSRA.S16 q0,q0,#16
 _NEON2SSE_INLINE int16x8_t vsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c) // VSRA.S16 q0,q0,#16
 {
     int16x8_t shift;
@@ -8020,7 +8022,7 @@ _NEON2SSE_INLINE int16x8_t vsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,
     return vaddq_s16(a, shift);
 }
 
-int32x4_t vsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VSRA.S32 q0,q0,#32
+_NEON2SSESTORAGE int32x4_t vsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VSRA.S32 q0,q0,#32
 _NEON2SSE_INLINE int32x4_t vsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c) // VSRA.S32 q0,q0,#32
 {
     int32x4_t shift;
@@ -8028,7 +8030,7 @@ _NEON2SSE_INLINE int32x4_t vsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,
     return vaddq_s32(a, shift);
 }
 
-int64x2_t vsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VSRA.S64 q0,q0,#64
+_NEON2SSESTORAGE int64x2_t vsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VSRA.S64 q0,q0,#64
 _NEON2SSE_INLINE int64x2_t vsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c) // VSRA.S64 q0,q0,#64
 {
     int64x2_t shift;
@@ -8036,7 +8038,7 @@ _NEON2SSE_INLINE int64x2_t vsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,
     return vaddq_s64( a, shift);
 }
 
-uint8x16_t vsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VSRA.U8 q0,q0,#8
+_NEON2SSESTORAGE uint8x16_t vsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VSRA.U8 q0,q0,#8
 _NEON2SSE_INLINE uint8x16_t vsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c) // VSRA.U8 q0,q0,#8
 {
     uint8x16_t shift;
@@ -8044,7 +8046,7 @@ _NEON2SSE_INLINE uint8x16_t vsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(
     return vaddq_u8(a, shift);
 }
 
-uint16x8_t vsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VSRA.s16 q0,q0,#16
+_NEON2SSESTORAGE uint16x8_t vsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VSRA.s16 q0,q0,#16
 _NEON2SSE_INLINE uint16x8_t vsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c) // VSRA.s16 q0,q0,#16
 {
     uint16x8_t shift;
@@ -8052,7 +8054,7 @@ _NEON2SSE_INLINE uint16x8_t vsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrange
     return vaddq_u16(a,  shift);
 }
 
-uint32x4_t vsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VSRA.U32 q0,q0,#32
+_NEON2SSESTORAGE uint32x4_t vsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VSRA.U32 q0,q0,#32
 _NEON2SSE_INLINE uint32x4_t vsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c) // VSRA.U32 q0,q0,#32
 {
     uint32x4_t shift;
@@ -8060,7 +8062,7 @@ _NEON2SSE_INLINE uint32x4_t vsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrange
     return vaddq_u32(a, shift);
 }
 
-uint64x2_t vsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VSRA.U64 q0,q0,#64
+_NEON2SSESTORAGE uint64x2_t vsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VSRA.U64 q0,q0,#64
 _NEON2SSE_INLINE uint64x2_t vsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c) // VSRA.U64 q0,q0,#64
 {
     uint64x2_t shift;
@@ -8070,7 +8072,7 @@ _NEON2SSE_INLINE uint64x2_t vsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrange
 
 //************* Vector rounding shift right by constant and accumulate ****************************
 //************************************************************************************************
-int8x8_t vrsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c); // VRSRA.S8 d0,d0,#8
+_NEON2SSESTORAGE int8x8_t vrsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c); // VRSRA.S8 d0,d0,#8
 _NEON2SSE_INLINE int8x8_t vrsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) int c) // VRSRA.S8 d0,d0,#8
 {
     int8x8_t shift;
@@ -8078,7 +8080,7 @@ _NEON2SSE_INLINE int8x8_t vrsra_n_s8(int8x8_t a, int8x8_t b, __constrange(1,8) i
     return vadd_s8( a, shift);
 }
 
-int16x4_t vrsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c); // VRSRA.S16 d0,d0,#16
+_NEON2SSESTORAGE int16x4_t vrsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c); // VRSRA.S16 d0,d0,#16
 _NEON2SSE_INLINE int16x4_t vrsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,16) int c) // VRSRA.S16 d0,d0,#16
 {
     int16x4_t shift;
@@ -8086,7 +8088,7 @@ _NEON2SSE_INLINE int16x4_t vrsra_n_s16(int16x4_t a, int16x4_t b, __constrange(1,
     return vadd_s16(a, shift);
 }
 
-int32x2_t vrsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c); // VRSRA.S32 d0,d0,#32
+_NEON2SSESTORAGE int32x2_t vrsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c); // VRSRA.S32 d0,d0,#32
 _NEON2SSE_INLINE int32x2_t vrsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,32) int c) // VRSRA.S32 d0,d0,#32
 {
     //may be not optimal compared with the serial execution
@@ -8095,7 +8097,7 @@ _NEON2SSE_INLINE int32x2_t vrsra_n_s32(int32x2_t a, int32x2_t b, __constrange(1,
     return vadd_s32( a, shift);
 }
 
-int64x1_t vrsra_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VRSRA.S64 d0,d0,#64
+_NEON2SSESTORAGE int64x1_t vrsra_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VRSRA.S64 d0,d0,#64
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vrsra_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c), _NEON2SSE_REASON_SLOW_SERIAL) //serial solution
 {
     int64x1_t shift;
@@ -8103,7 +8105,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vrsra_n_s64(int64x1_t a
     return vadd_s64( a, shift);
 }
 
-uint8x8_t vrsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c); // VRSRA.U8 d0,d0,#8
+_NEON2SSESTORAGE uint8x8_t vrsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c); // VRSRA.U8 d0,d0,#8
 _NEON2SSE_INLINE uint8x8_t vrsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8) int c) // VRSRA.U8 d0,d0,#8
 {
     uint8x8_t shift;
@@ -8111,7 +8113,7 @@ _NEON2SSE_INLINE uint8x8_t vrsra_n_u8(uint8x8_t a, uint8x8_t b, __constrange(1,8
     return vadd_u8(a, shift);
 }
 
-uint16x4_t vrsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c); // VRSRA.s16 d0,d0,#16
+_NEON2SSESTORAGE uint16x4_t vrsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c); // VRSRA.s16 d0,d0,#16
 _NEON2SSE_INLINE uint16x4_t vrsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange(1,16) int c) // VRSRA.s16 d0,d0,#16
 {
     uint16x4_t shift;
@@ -8119,7 +8121,7 @@ _NEON2SSE_INLINE uint16x4_t vrsra_n_u16(uint16x4_t a, uint16x4_t b, __constrange
     return vadd_u16(a,shift);
 }
 
-uint32x2_t vrsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c); // VRSRA.U32 d0,d0,#32
+_NEON2SSESTORAGE uint32x2_t vrsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c); // VRSRA.U32 d0,d0,#32
 _NEON2SSE_INLINE uint32x2_t vrsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange(1,32) int c) // VRSRA.U32 d0,d0,#32
 {
     //may be not optimal compared with the serial execution
@@ -8128,7 +8130,7 @@ _NEON2SSE_INLINE uint32x2_t vrsra_n_u32(uint32x2_t a, uint32x2_t b, __constrange
     return vadd_u32( a, shift);
 }
 
-uint64x1_t vrsra_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VRSRA.U64 d0,d0,#64
+_NEON2SSESTORAGE uint64x1_t vrsra_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VRSRA.U64 d0,d0,#64
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vrsra_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c), _NEON2SSE_REASON_SLOW_SERIAL) //serial solution
 {
     //may be not optimal compared with the serial execution
@@ -8137,7 +8139,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vrsra_n_u64(uint64x1_t
     return vadd_u64( a, shift);
 }
 
-int8x16_t vrsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VRSRA.S8 q0,q0,#8
+_NEON2SSESTORAGE int8x16_t vrsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VRSRA.S8 q0,q0,#8
 _NEON2SSE_INLINE int8x16_t vrsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c) // VRSRA.S8 q0,q0,#8
 {
     int8x16_t shift;
@@ -8145,7 +8147,7 @@ _NEON2SSE_INLINE int8x16_t vrsraq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,
     return vaddq_s8(a, shift);
 }
 
-int16x8_t vrsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VRSRA.S16 q0,q0,#16
+_NEON2SSESTORAGE int16x8_t vrsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VRSRA.S16 q0,q0,#16
 _NEON2SSE_INLINE int16x8_t vrsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c) // VRSRA.S16 q0,q0,#16
 {
     int16x8_t shift;
@@ -8153,7 +8155,7 @@ _NEON2SSE_INLINE int16x8_t vrsraq_n_s16(int16x8_t a, int16x8_t b, __constrange(1
     return vaddq_s16(a, shift);
 }
 
-int32x4_t vrsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VRSRA.S32 q0,q0,#32
+_NEON2SSESTORAGE int32x4_t vrsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VRSRA.S32 q0,q0,#32
 _NEON2SSE_INLINE int32x4_t vrsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c) // VRSRA.S32 q0,q0,#32
 {
     int32x4_t shift;
@@ -8161,7 +8163,7 @@ _NEON2SSE_INLINE int32x4_t vrsraq_n_s32(int32x4_t a, int32x4_t b, __constrange(1
     return vaddq_s32(a, shift);
 }
 
-int64x2_t vrsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VRSRA.S64 q0,q0,#64
+_NEON2SSESTORAGE int64x2_t vrsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VRSRA.S64 q0,q0,#64
 _NEON2SSE_INLINE int64x2_t vrsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c)
 {
     int64x2_t shift;
@@ -8169,7 +8171,7 @@ _NEON2SSE_INLINE int64x2_t vrsraq_n_s64(int64x2_t a, int64x2_t b, __constrange(1
     return vaddq_s64(a, shift);
 }
 
-uint8x16_t vrsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VRSRA.U8 q0,q0,#8
+_NEON2SSESTORAGE uint8x16_t vrsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VRSRA.U8 q0,q0,#8
 _NEON2SSE_INLINE uint8x16_t vrsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c) // VRSRA.U8 q0,q0,#8
 {
     uint8x16_t shift;
@@ -8177,7 +8179,7 @@ _NEON2SSE_INLINE uint8x16_t vrsraq_n_u8(uint8x16_t a, uint8x16_t b, __constrange
     return vaddq_u8(a, shift);
 }
 
-uint16x8_t vrsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VRSRA.s16 q0,q0,#16
+_NEON2SSESTORAGE uint16x8_t vrsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VRSRA.s16 q0,q0,#16
 _NEON2SSE_INLINE uint16x8_t vrsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c) // VRSRA.s16 q0,q0,#16
 {
     uint16x8_t shift;
@@ -8185,7 +8187,7 @@ _NEON2SSE_INLINE uint16x8_t vrsraq_n_u16(uint16x8_t a, uint16x8_t b, __constrang
     return vaddq_u16(a,  shift);
 }
 
-uint32x4_t vrsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VRSRA.U32 q0,q0,#32
+_NEON2SSESTORAGE uint32x4_t vrsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VRSRA.U32 q0,q0,#32
 _NEON2SSE_INLINE uint32x4_t vrsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c) // VRSRA.U32 q0,q0,#32
 {
     uint32x4_t shift;
@@ -8193,7 +8195,7 @@ _NEON2SSE_INLINE uint32x4_t vrsraq_n_u32(uint32x4_t a, uint32x4_t b, __constrang
     return vaddq_u32(a, shift);
 }
 
-uint64x2_t vrsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VRSRA.U64 q0,q0,#64
+_NEON2SSESTORAGE uint64x2_t vrsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VRSRA.U64 q0,q0,#64
 _NEON2SSE_INLINE uint64x2_t vrsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c)
 {
     uint64x2_t shift;
@@ -8204,7 +8206,7 @@ _NEON2SSE_INLINE uint64x2_t vrsraq_n_u64(uint64x2_t a, uint64x2_t b, __constrang
 //**********************Vector saturating shift left by constant *****************************
 //********************************************************************************************
 //we don't check const ranges  assuming they are met
-int8x8_t vqshl_n_s8(int8x8_t a, __constrange(0,7) int b); // VQSHL.S8 d0,d0,#0
+_NEON2SSESTORAGE int8x8_t vqshl_n_s8(int8x8_t a, __constrange(0,7) int b); // VQSHL.S8 d0,d0,#0
 _NEON2SSE_INLINE int8x8_t vqshl_n_s8(int8x8_t a, __constrange(0,7) int b) // VQSHL.S8 d0,d0,#0
 {
     //no 8 bit shift available in IA32 SIMD, go to 16 bit. It also provides the auto saturation (in packs function)
@@ -8216,7 +8218,7 @@ _NEON2SSE_INLINE int8x8_t vqshl_n_s8(int8x8_t a, __constrange(0,7) int b) // VQS
     return64(r128);
 }
 
-int16x4_t vqshl_n_s16(int16x4_t a, __constrange(0,15) int b); // VQSHL.S16 d0,d0,#0
+_NEON2SSESTORAGE int16x4_t vqshl_n_s16(int16x4_t a, __constrange(0,15) int b); // VQSHL.S16 d0,d0,#0
 _NEON2SSE_INLINE int16x4_t vqshl_n_s16(int16x4_t a, __constrange(0,15) int b) // VQSHL.S16 d0,d0,#0
 {
     // go to 32 bit to get the auto saturation (in packs function)
@@ -8228,7 +8230,7 @@ _NEON2SSE_INLINE int16x4_t vqshl_n_s16(int16x4_t a, __constrange(0,15) int b) //
     return64(r128);
 }
 
-int32x2_t vqshl_n_s32(int32x2_t a,  __constrange(0,31) int b); // VQSHL.S32 d0,d0,#0
+_NEON2SSESTORAGE int32x2_t vqshl_n_s32(int32x2_t a,  __constrange(0,31) int b); // VQSHL.S32 d0,d0,#0
 _NEON2SSE_INLINE int32x2_t vqshl_n_s32(int32x2_t a,  __constrange(0,31) int b)
 {
     //serial execution may be faster
@@ -8237,7 +8239,7 @@ _NEON2SSE_INLINE int32x2_t vqshl_n_s32(int32x2_t a,  __constrange(0,31) int b)
 }
 
 
-int64x1_t vqshl_n_s64(int64x1_t a, __constrange(0,63) int b); // VQSHL.S64 d0,d0,#0
+_NEON2SSESTORAGE int64x1_t vqshl_n_s64(int64x1_t a, __constrange(0,63) int b); // VQSHL.S64 d0,d0,#0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vqshl_n_s64(int64x1_t a, __constrange(0,63) int b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     // no effective SIMD solution here
@@ -8254,7 +8256,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x1_t vqshl_n_s64(int64x1_t a
 }
 
 
-uint8x8_t vqshl_n_u8(uint8x8_t a, __constrange(0,7) int b); // VQSHL.U8 d0,d0,#0
+_NEON2SSESTORAGE uint8x8_t vqshl_n_u8(uint8x8_t a, __constrange(0,7) int b); // VQSHL.U8 d0,d0,#0
 _NEON2SSE_INLINE uint8x8_t vqshl_n_u8(uint8x8_t a, __constrange(0,7) int b) // VQSHL.U8 d0,d0,#0
 {
     //no 8 bit shift available in IA32 SIMD, go to 16 bit
@@ -8266,7 +8268,7 @@ _NEON2SSE_INLINE uint8x8_t vqshl_n_u8(uint8x8_t a, __constrange(0,7) int b) // V
     return64(r128);
 }
 
-uint16x4_t vqshl_n_u16(uint16x4_t a, __constrange(0,15) int b); // VQSHL.s16 d0,d0,#0
+_NEON2SSESTORAGE uint16x4_t vqshl_n_u16(uint16x4_t a, __constrange(0,15) int b); // VQSHL.s16 d0,d0,#0
 _NEON2SSE_INLINE uint16x4_t vqshl_n_u16(uint16x4_t a, __constrange(0,15) int b) // VQSHL.s16 d0,d0,#0
 {
     // go to 32 bit to get the auto saturation (in packus function)
@@ -8278,14 +8280,14 @@ _NEON2SSE_INLINE uint16x4_t vqshl_n_u16(uint16x4_t a, __constrange(0,15) int b)
     return64(r128);
 }
 
-uint32x2_t vqshl_n_u32(uint32x2_t a,  __constrange(0,31) int b); // VQSHL.U32 d0,d0,#0
+_NEON2SSESTORAGE uint32x2_t vqshl_n_u32(uint32x2_t a,  __constrange(0,31) int b); // VQSHL.U32 d0,d0,#0
 _NEON2SSE_INLINE uint32x2_t vqshl_n_u32(uint32x2_t a,  __constrange(0,31) int b)
 {
     uint32x2_t res64;
     return64(vqshlq_n_u32(_pM128i(a), b));
 }
 
-uint64x1_t vqshl_n_u64(uint64x1_t a, __constrange(0,63) int b); // VQSHL.U64 d0,d0,#0
+_NEON2SSESTORAGE uint64x1_t vqshl_n_u64(uint64x1_t a, __constrange(0,63) int b); // VQSHL.U64 d0,d0,#0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vqshl_n_u64(uint64x1_t a, __constrange(0,63) int b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     // no effective SIMD solution here
@@ -8297,7 +8299,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vqshl_n_u64(uint64x1_t
     return res;
 }
 
-int8x16_t vqshlq_n_s8(int8x16_t a, __constrange(0,7) int b); // VQSHL.S8 q0,q0,#0
+_NEON2SSESTORAGE int8x16_t vqshlq_n_s8(int8x16_t a, __constrange(0,7) int b); // VQSHL.S8 q0,q0,#0
 _NEON2SSE_INLINE int8x16_t vqshlq_n_s8(int8x16_t a, __constrange(0,7) int b) // VQSHL.S8 q0,q0,#0
 {
     // go to 16 bit to get the auto saturation (in packs function)
@@ -8311,7 +8313,7 @@ _NEON2SSE_INLINE int8x16_t vqshlq_n_s8(int8x16_t a, __constrange(0,7) int b) //
     return _mm_packs_epi16 (r128_1, r128_2); //saturated s8
 }
 
-int16x8_t vqshlq_n_s16(int16x8_t a, __constrange(0,15) int b); // VQSHL.S16 q0,q0,#0
+_NEON2SSESTORAGE int16x8_t vqshlq_n_s16(int16x8_t a, __constrange(0,15) int b); // VQSHL.S16 q0,q0,#0
 _NEON2SSE_INLINE int16x8_t vqshlq_n_s16(int16x8_t a, __constrange(0,15) int b) // VQSHL.S16 q0,q0,#0
 {
     // manual saturation solution looks LESS optimal than 32 bits conversion one
@@ -8326,7 +8328,7 @@ _NEON2SSE_INLINE int16x8_t vqshlq_n_s16(int16x8_t a, __constrange(0,15) int b) /
     return _mm_packs_epi32 (r128_1, r128_2); //saturated s16
 }
 
-int32x4_t vqshlq_n_s32(int32x4_t a, __constrange(0,31) int b); // VQSHL.S32 q0,q0,#0
+_NEON2SSESTORAGE int32x4_t vqshlq_n_s32(int32x4_t a, __constrange(0,31) int b); // VQSHL.S32 q0,q0,#0
 _NEON2SSE_INLINE int32x4_t vqshlq_n_s32(int32x4_t a, __constrange(0,31) int b) // VQSHL.S32 q0,q0,#0
 {
     // no 64 bit saturation option available, special tricks necessary
@@ -8347,7 +8349,7 @@ _NEON2SSE_INLINE int32x4_t vqshlq_n_s32(int32x4_t a, __constrange(0,31) int b) /
     return _mm_or_si128 (c7ffffff_mask, shift_res_mask);
 }
 
-int64x2_t vqshlq_n_s64(int64x2_t a, __constrange(0,63) int b); // VQSHL.S64 q0,q0,#0
+_NEON2SSESTORAGE int64x2_t vqshlq_n_s64(int64x2_t a, __constrange(0,63) int b); // VQSHL.S64 q0,q0,#0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqshlq_n_s64(int64x2_t a, __constrange(0,63) int b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     // no effective SIMD solution here
@@ -8366,7 +8368,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqshlq_n_s64(int64x2_t
     return _mm_load_si128((__m128i*)res);
 }
 
-uint8x16_t vqshlq_n_u8(uint8x16_t a, __constrange(0,7) int b); // VQSHL.U8 q0,q0,#0
+_NEON2SSESTORAGE uint8x16_t vqshlq_n_u8(uint8x16_t a, __constrange(0,7) int b); // VQSHL.U8 q0,q0,#0
 _NEON2SSE_INLINE uint8x16_t vqshlq_n_u8(uint8x16_t a, __constrange(0,7) int b) // VQSHL.U8 q0,q0,#0
 {
     // go to 16 bit to get the auto saturation (in packs function)
@@ -8380,7 +8382,7 @@ _NEON2SSE_INLINE uint8x16_t vqshlq_n_u8(uint8x16_t a, __constrange(0,7) int b) /
     return _mm_packus_epi16 (r128_1, r128_2); //saturated u8
 }
 
-uint16x8_t vqshlq_n_u16(uint16x8_t a, __constrange(0,15) int b); // VQSHL.s16 q0,q0,#0
+_NEON2SSESTORAGE uint16x8_t vqshlq_n_u16(uint16x8_t a, __constrange(0,15) int b); // VQSHL.s16 q0,q0,#0
 _NEON2SSE_INLINE uint16x8_t vqshlq_n_u16(uint16x8_t a, __constrange(0,15) int b) // VQSHL.s16 q0,q0,#0
 {
     // manual saturation solution looks more optimal than 32 bits conversion one
@@ -8394,7 +8396,7 @@ _NEON2SSE_INLINE uint16x8_t vqshlq_n_u16(uint16x8_t a, __constrange(0,15) int b)
     return _mm_or_si128 (shift_res, saturation_mask);
 }
 
-uint32x4_t vqshlq_n_u32(uint32x4_t a, __constrange(0,31) int b); // VQSHL.U32 q0,q0,#0
+_NEON2SSESTORAGE uint32x4_t vqshlq_n_u32(uint32x4_t a, __constrange(0,31) int b); // VQSHL.U32 q0,q0,#0
 _NEON2SSE_INLINE uint32x4_t vqshlq_n_u32(uint32x4_t a, __constrange(0,31) int b) // VQSHL.U32 q0,q0,#0
 {
     // manual saturation solution, no 64 bit saturation option, the serial version may be faster
@@ -8408,7 +8410,7 @@ _NEON2SSE_INLINE uint32x4_t vqshlq_n_u32(uint32x4_t a, __constrange(0,31) int b)
     return _mm_or_si128 (shift_res, saturation_mask);
 }
 
-uint64x2_t vqshlq_n_u64(uint64x2_t a, __constrange(0,63) int b); // VQSHL.U64 q0,q0,#0
+_NEON2SSESTORAGE uint64x2_t vqshlq_n_u64(uint64x2_t a, __constrange(0,63) int b); // VQSHL.U64 q0,q0,#0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vqshlq_n_u64(uint64x2_t a, __constrange(0,63) int b), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     // no effective SIMD solution here
@@ -8425,7 +8427,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vqshlq_n_u64(uint64x2_
 
 //**************Vector signed->unsigned saturating shift left by constant *************
 //*************************************************************************************
-uint8x8_t vqshlu_n_s8(int8x8_t a, __constrange(0,7) int b); // VQSHLU.S8 d0,d0,#0
+_NEON2SSESTORAGE uint8x8_t vqshlu_n_s8(int8x8_t a, __constrange(0,7) int b); // VQSHLU.S8 d0,d0,#0
 _NEON2SSE_INLINE uint8x8_t vqshlu_n_s8(int8x8_t a, __constrange(0,7) int b) // VQSHLU.S8 d0,d0,#0
 {
     //no 8 bit shift available in IA32 SIMD, go to 16 bit. It also provides the auto saturation (in packs function)
@@ -8437,7 +8439,7 @@ _NEON2SSE_INLINE uint8x8_t vqshlu_n_s8(int8x8_t a, __constrange(0,7) int b) // V
     return64(r128);
 }
 
-uint16x4_t vqshlu_n_s16(int16x4_t a, __constrange(0,15) int b); // VQSHLU.S16 d0,d0,#0
+_NEON2SSESTORAGE uint16x4_t vqshlu_n_s16(int16x4_t a, __constrange(0,15) int b); // VQSHLU.S16 d0,d0,#0
 _NEON2SSE_INLINE uint16x4_t vqshlu_n_s16(int16x4_t a, __constrange(0,15) int b) // VQSHLU.S16 d0,d0,#0
 {
     uint16x4_t res64;
@@ -8448,14 +8450,14 @@ _NEON2SSE_INLINE uint16x4_t vqshlu_n_s16(int16x4_t a, __constrange(0,15) int b)
     return64(r128);
 }
 
-uint32x2_t vqshlu_n_s32(int32x2_t a,  __constrange(0,31) int b); // VQSHLU.S32 d0,d0,#0
+_NEON2SSESTORAGE uint32x2_t vqshlu_n_s32(int32x2_t a,  __constrange(0,31) int b); // VQSHLU.S32 d0,d0,#0
 _NEON2SSE_INLINE int32x2_t vqshlu_n_s32(int32x2_t a,  __constrange(0,31) int b)
 {
     int32x2_t res64;
     return64( vqshluq_n_s32(_pM128i(a), b));
 }
 
-uint64x1_t vqshlu_n_s64(int64x1_t a, __constrange(0,63) int b); // VQSHLU.S64 d0,d0,#0
+_NEON2SSESTORAGE uint64x1_t vqshlu_n_s64(int64x1_t a, __constrange(0,63) int b); // VQSHLU.S64 d0,d0,#0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vqshlu_n_s64(int64x1_t a, __constrange(0,63) int b), _NEON2SSE_REASON_SLOW_SERIAL) // no effective SIMD solution here, serial execution looks faster
 {
     uint64x1_t res;
@@ -8469,7 +8471,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x1_t vqshlu_n_s64(int64x1_t
     return res;
 }
 
-uint8x16_t vqshluq_n_s8(int8x16_t a, __constrange(0,7) int b); // VQSHLU.S8 q0,q0,#0
+_NEON2SSESTORAGE uint8x16_t vqshluq_n_s8(int8x16_t a, __constrange(0,7) int b); // VQSHLU.S8 q0,q0,#0
 _NEON2SSE_INLINE uint8x16_t vqshluq_n_s8(int8x16_t a, __constrange(0,7) int b) // VQSHLU.S8 q0,q0,#0
 {
     __m128i a128, r128_1, r128_2;
@@ -8482,7 +8484,7 @@ _NEON2SSE_INLINE uint8x16_t vqshluq_n_s8(int8x16_t a, __constrange(0,7) int b) /
     return _mm_packus_epi16 (r128_1, r128_2); //saturated u8
 }
 
-uint16x8_t vqshluq_n_s16(int16x8_t a, __constrange(0,15) int b); // VQSHLU.S16 q0,q0,#0
+_NEON2SSESTORAGE uint16x8_t vqshluq_n_s16(int16x8_t a, __constrange(0,15) int b); // VQSHLU.S16 q0,q0,#0
 _NEON2SSE_INLINE uint16x8_t vqshluq_n_s16(int16x8_t a, __constrange(0,15) int b) // VQSHLU.S16 q0,q0,#0
 {
     // manual saturation solution looks LESS optimal than 32 bits conversion one
@@ -8496,7 +8498,7 @@ _NEON2SSE_INLINE uint16x8_t vqshluq_n_s16(int16x8_t a, __constrange(0,15) int b)
     return _MM_PACKUS_EPI32 (r128_1, r128_2); //saturated s16
 }
 
-uint32x4_t vqshluq_n_s32(int32x4_t a, __constrange(0,31) int b); // VQSHLU.S32 q0,q0,#0
+_NEON2SSESTORAGE uint32x4_t vqshluq_n_s32(int32x4_t a, __constrange(0,31) int b); // VQSHLU.S32 q0,q0,#0
 _NEON2SSE_INLINE uint32x4_t vqshluq_n_s32(int32x4_t a, __constrange(0,31) int b) // VQSHLU.S32 q0,q0,#0
 {
     //solution may be  not optimal compared with the serial one
@@ -8514,7 +8516,7 @@ _NEON2SSE_INLINE uint32x4_t vqshluq_n_s32(int32x4_t a, __constrange(0,31) int b)
     return _mm_or_si128 (a_shift, a_masked); //actual saturation
 }
 
-uint64x2_t vqshluq_n_s64(int64x2_t a, __constrange(0,63) int b); // VQSHLU.S64 q0,q0,#0
+_NEON2SSESTORAGE uint64x2_t vqshluq_n_s64(int64x2_t a, __constrange(0,63) int b); // VQSHLU.S64 q0,q0,#0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vqshluq_n_s64(int64x2_t a, __constrange(0,63) int b),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     // no effective SIMD solution here, serial execution looks faster
@@ -8536,7 +8538,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint64x2_t vqshluq_n_s64(int64x2_
 
 //************** Vector narrowing  shift right by constant **************
 //**********************************************************************
-int8x8_t vshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VSHRN.I16 d0,q0,#8
+_NEON2SSESTORAGE int8x8_t vshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VSHRN.I16 d0,q0,#8
 _NEON2SSE_INLINE int8x8_t vshrn_n_s16(int16x8_t a, __constrange(1,8) int b) // VSHRN.I16 d0,q0,#8
 {
     int8x8_t res64;
@@ -8546,7 +8548,7 @@ _NEON2SSE_INLINE int8x8_t vshrn_n_s16(int16x8_t a, __constrange(1,8) int b) // V
     return64(r16);
 }
 
-int16x4_t vshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VSHRN.I32 d0,q0,#16
+_NEON2SSESTORAGE int16x4_t vshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VSHRN.I32 d0,q0,#16
 _NEON2SSE_INLINE int16x4_t vshrn_n_s32(int32x4_t a, __constrange(1,16) int b) // VSHRN.I32 d0,q0,#16
 {
     int16x4_t res64;
@@ -8556,7 +8558,7 @@ _NEON2SSE_INLINE int16x4_t vshrn_n_s32(int32x4_t a, __constrange(1,16) int b) //
     return64(r32);
 }
 
-int32x2_t vshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VSHRN.I64 d0,q0,#32
+_NEON2SSESTORAGE int32x2_t vshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VSHRN.I64 d0,q0,#32
 _NEON2SSE_INLINE int32x2_t vshrn_n_s64(int64x2_t a, __constrange(1,32) int b)
 {
     int32x2_t res64;
@@ -8566,7 +8568,7 @@ _NEON2SSE_INLINE int32x2_t vshrn_n_s64(int64x2_t a, __constrange(1,32) int b)
     return64(r64);
 }
 
-uint8x8_t vshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VSHRN.I16 d0,q0,#8
+_NEON2SSESTORAGE uint8x8_t vshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VSHRN.I16 d0,q0,#8
 _NEON2SSE_INLINE uint8x8_t vshrn_n_u16(uint16x8_t a, __constrange(1,8) int b) // VSHRN.I16 d0,q0,#8
 {
     uint8x8_t res64;
@@ -8578,7 +8580,7 @@ _NEON2SSE_INLINE uint8x8_t vshrn_n_u16(uint16x8_t a, __constrange(1,8) int b) //
     return64(r16);
 }
 
-uint16x4_t vshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VSHRN.I32 d0,q0,#16
+_NEON2SSESTORAGE uint16x4_t vshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VSHRN.I32 d0,q0,#16
 _NEON2SSE_INLINE uint16x4_t vshrn_n_u32(uint32x4_t a, __constrange(1,16) int b) // VSHRN.I32 d0,q0,#16
 {
     uint16x4_t res64;
@@ -8590,7 +8592,7 @@ _NEON2SSE_INLINE uint16x4_t vshrn_n_u32(uint32x4_t a, __constrange(1,16) int b)
     return64(r32);
 }
 
-uint32x2_t vshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VSHRN.I64 d0,q0,#32
+_NEON2SSESTORAGE uint32x2_t vshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VSHRN.I64 d0,q0,#32
 _NEON2SSE_INLINE uint32x2_t vshrn_n_u64(uint64x2_t a, __constrange(1,32) int b)
 {
     uint32x2_t res64;
@@ -8602,7 +8604,7 @@ _NEON2SSE_INLINE uint32x2_t vshrn_n_u64(uint64x2_t a, __constrange(1,32) int b)
 
 //************** Vector signed->unsigned narrowing saturating shift right by constant ********
 //*********************************************************************************************
-uint8x8_t vqshrun_n_s16(int16x8_t a, __constrange(1,8) int b); // VQSHRUN.S16 d0,q0,#8
+_NEON2SSESTORAGE uint8x8_t vqshrun_n_s16(int16x8_t a, __constrange(1,8) int b); // VQSHRUN.S16 d0,q0,#8
 _NEON2SSE_INLINE uint8x8_t vqshrun_n_s16(int16x8_t a, __constrange(1,8) int b) // VQSHRUN.S16 d0,q0,#8
 {
     uint8x8_t res64;
@@ -8612,7 +8614,7 @@ _NEON2SSE_INLINE uint8x8_t vqshrun_n_s16(int16x8_t a, __constrange(1,8) int b) /
     return64(r16);
 }
 
-uint16x4_t vqshrun_n_s32(int32x4_t a, __constrange(1,16) int b); // VQSHRUN.S32 d0,q0,#16
+_NEON2SSESTORAGE uint16x4_t vqshrun_n_s32(int32x4_t a, __constrange(1,16) int b); // VQSHRUN.S32 d0,q0,#16
 _NEON2SSE_INLINE uint16x4_t vqshrun_n_s32(int32x4_t a, __constrange(1,16) int b) // VQSHRUN.S32 d0,q0,#16
 {
     uint16x4_t res64;
@@ -8622,7 +8624,7 @@ _NEON2SSE_INLINE uint16x4_t vqshrun_n_s32(int32x4_t a, __constrange(1,16) int b)
     return64(r32);
 }
 
-uint32x2_t vqshrun_n_s64(int64x2_t a, __constrange(1,32) int b); // VQSHRUN.S64 d0,q0,#32
+_NEON2SSESTORAGE uint32x2_t vqshrun_n_s64(int64x2_t a, __constrange(1,32) int b); // VQSHRUN.S64 d0,q0,#32
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vqshrun_n_s64(int64x2_t a, __constrange(1,32) int b), _NEON2SSE_REASON_SLOW_SERIAL) //serial solution is faster
 {
     _NEON2SSE_ALIGN_16 int64_t atmp[2];
@@ -8645,7 +8647,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vqshrun_n_s64(int64x2_
 }
 
 //**** Vector signed->unsigned rounding narrowing saturating shift right by constant *****
-uint8x8_t vqrshrun_n_s16(int16x8_t a, __constrange(1,8) int b); // VQRSHRUN.S16 d0,q0,#8
+_NEON2SSESTORAGE uint8x8_t vqrshrun_n_s16(int16x8_t a, __constrange(1,8) int b); // VQRSHRUN.S16 d0,q0,#8
 _NEON2SSE_INLINE uint8x8_t vqrshrun_n_s16(int16x8_t a, __constrange(1,8) int b) // VQRSHRUN.S16 d0,q0,#8
 {
     //solution may be not optimal compared with the serial one
@@ -8656,7 +8658,7 @@ _NEON2SSE_INLINE uint8x8_t vqrshrun_n_s16(int16x8_t a, __constrange(1,8) int b)
     return64(r16);
 }
 
-uint16x4_t vqrshrun_n_s32(int32x4_t a, __constrange(1,16) int b); // VQRSHRUN.S32 d0,q0,#16
+_NEON2SSESTORAGE uint16x4_t vqrshrun_n_s32(int32x4_t a, __constrange(1,16) int b); // VQRSHRUN.S32 d0,q0,#16
 _NEON2SSE_INLINE uint16x4_t vqrshrun_n_s32(int32x4_t a, __constrange(1,16) int b) // VQRSHRUN.S32 d0,q0,#16
 {
     //solution may be not optimal compared with the serial one
@@ -8667,7 +8669,7 @@ _NEON2SSE_INLINE uint16x4_t vqrshrun_n_s32(int32x4_t a, __constrange(1,16) int b
     return64(r32);
 }
 
-uint32x2_t vqrshrun_n_s64(int64x2_t a, __constrange(1,32) int b); // VQRSHRUN.S64 d0,q0,#32
+_NEON2SSESTORAGE uint32x2_t vqrshrun_n_s64(int64x2_t a, __constrange(1,32) int b); // VQRSHRUN.S64 d0,q0,#32
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vqrshrun_n_s64(int64x2_t a, __constrange(1,32) int b), _NEON2SSE_REASON_SLOW_SERIAL) //serial solution is faster
 {
     _NEON2SSE_ALIGN_16 int64_t atmp[2];
@@ -8691,7 +8693,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vqrshrun_n_s64(int64x2
 
 //***** Vector narrowing saturating shift right by constant ******
 //*****************************************************************
-int8x8_t vqshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VQSHRN.S16 d0,q0,#8
+_NEON2SSESTORAGE int8x8_t vqshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VQSHRN.S16 d0,q0,#8
 _NEON2SSE_INLINE int8x8_t vqshrn_n_s16(int16x8_t a, __constrange(1,8) int b) // VQSHRN.S16 d0,q0,#8
 {
     int8x8_t res64;
@@ -8701,7 +8703,7 @@ _NEON2SSE_INLINE int8x8_t vqshrn_n_s16(int16x8_t a, __constrange(1,8) int b) //
     return64(r16);
 }
 
-int16x4_t vqshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VQSHRN.S32 d0,q0,#16
+_NEON2SSESTORAGE int16x4_t vqshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VQSHRN.S32 d0,q0,#16
 _NEON2SSE_INLINE int16x4_t vqshrn_n_s32(int32x4_t a, __constrange(1,16) int b) // VQSHRN.S32 d0,q0,#16
 {
     int16x4_t res64;
@@ -8711,7 +8713,7 @@ _NEON2SSE_INLINE int16x4_t vqshrn_n_s32(int32x4_t a, __constrange(1,16) int b) /
     return64(r32);
 }
 
-int32x2_t vqshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VQSHRN.S64 d0,q0,#32
+_NEON2SSESTORAGE int32x2_t vqshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VQSHRN.S64 d0,q0,#32
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqshrn_n_s64(int64x2_t a, __constrange(1,32) int b), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     //no optimal SIMD solution found
@@ -8729,7 +8731,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqshrn_n_s64(int64x2_t
     return res;
 }
 
-uint8x8_t vqshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VQSHRN.s16 d0,q0,#8
+_NEON2SSESTORAGE uint8x8_t vqshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VQSHRN.s16 d0,q0,#8
 _NEON2SSE_INLINE uint8x8_t vqshrn_n_u16(uint16x8_t a, __constrange(1,8) int b) // VQSHRN.s16 d0,q0,#8
 {
     uint8x8_t res64;
@@ -8739,7 +8741,7 @@ _NEON2SSE_INLINE uint8x8_t vqshrn_n_u16(uint16x8_t a, __constrange(1,8) int b) /
     return64(r16);
 }
 
-uint16x4_t vqshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VQSHRN.U32 d0,q0,#16
+_NEON2SSESTORAGE uint16x4_t vqshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VQSHRN.U32 d0,q0,#16
 _NEON2SSE_INLINE uint16x4_t vqshrn_n_u32(uint32x4_t a, __constrange(1,16) int b) // VQSHRN.U32 d0,q0,#16
 {
     uint16x4_t res64;
@@ -8749,7 +8751,7 @@ _NEON2SSE_INLINE uint16x4_t vqshrn_n_u32(uint32x4_t a, __constrange(1,16) int b)
     return64(r32);
 }
 
-uint32x2_t vqshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VQSHRN.U64 d0,q0,#32
+_NEON2SSESTORAGE uint32x2_t vqshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VQSHRN.U64 d0,q0,#32
 _NEON2SSE_INLINE uint32x2_t vqshrn_n_u64(uint64x2_t a, __constrange(1,32) int b)
 {
     //serial solution may be faster
@@ -8767,7 +8769,7 @@ _NEON2SSE_INLINE uint32x2_t vqshrn_n_u64(uint64x2_t a, __constrange(1,32) int b)
 
 //********* Vector rounding narrowing shift right by constant *************************
 //****************************************************************************************
-int8x8_t vrshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VRSHRN.I16 d0,q0,#8
+_NEON2SSESTORAGE int8x8_t vrshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VRSHRN.I16 d0,q0,#8
 _NEON2SSE_INLINE int8x8_t vrshrn_n_s16(int16x8_t a, __constrange(1,8) int b) // VRSHRN.I16 d0,q0,#8
 {
     int8x8_t res64;
@@ -8777,7 +8779,7 @@ _NEON2SSE_INLINE int8x8_t vrshrn_n_s16(int16x8_t a, __constrange(1,8) int b) //
     return64(r16);
 }
 
-int16x4_t vrshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VRSHRN.I32 d0,q0,#16
+_NEON2SSESTORAGE int16x4_t vrshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VRSHRN.I32 d0,q0,#16
 _NEON2SSE_INLINE int16x4_t vrshrn_n_s32(int32x4_t a, __constrange(1,16) int b) // VRSHRN.I32 d0,q0,#16
 {
     int16x4_t res64;
@@ -8787,7 +8789,7 @@ _NEON2SSE_INLINE int16x4_t vrshrn_n_s32(int32x4_t a, __constrange(1,16) int b) /
     return64(r32);
 }
 
-int32x2_t vrshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VRSHRN.I64 d0,q0,#32
+_NEON2SSESTORAGE int32x2_t vrshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VRSHRN.I64 d0,q0,#32
 _NEON2SSE_INLINE int32x2_t vrshrn_n_s64(int64x2_t a, __constrange(1,32) int b)
 {
     int32x2_t res64;
@@ -8797,7 +8799,7 @@ _NEON2SSE_INLINE int32x2_t vrshrn_n_s64(int64x2_t a, __constrange(1,32) int b)
     return64(r64);
 }
 
-uint8x8_t vrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VRSHRN.I16 d0,q0,#8
+_NEON2SSESTORAGE uint8x8_t vrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VRSHRN.I16 d0,q0,#8
 _NEON2SSE_INLINE uint8x8_t vrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b) // VRSHRN.I16 d0,q0,#8
 {
     uint8x8_t res64;
@@ -8809,7 +8811,7 @@ _NEON2SSE_INLINE uint8x8_t vrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b) /
     return64(r16);
 }
 
-uint16x4_t vrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VRSHRN.I32 d0,q0,#16
+_NEON2SSESTORAGE uint16x4_t vrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VRSHRN.I32 d0,q0,#16
 _NEON2SSE_INLINE uint16x4_t vrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b) // VRSHRN.I32 d0,q0,#16
 {
     uint16x4_t res64;
@@ -8821,7 +8823,7 @@ _NEON2SSE_INLINE uint16x4_t vrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b)
     return64(r32);
 }
 
-uint32x2_t vrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VRSHRN.I64 d0,q0,#32
+_NEON2SSESTORAGE uint32x2_t vrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VRSHRN.I64 d0,q0,#32
 _NEON2SSE_INLINE uint32x2_t vrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b) //serial solution may be faster
 {
     uint32x2_t res64;
@@ -8833,7 +8835,7 @@ _NEON2SSE_INLINE uint32x2_t vrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b)
 
 //************* Vector rounding narrowing saturating shift right by constant ************
 //****************************************************************************************
-int8x8_t vqrshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VQRSHRN.S16 d0,q0,#8
+_NEON2SSESTORAGE int8x8_t vqrshrn_n_s16(int16x8_t a, __constrange(1,8) int b); // VQRSHRN.S16 d0,q0,#8
 _NEON2SSE_INLINE int8x8_t vqrshrn_n_s16(int16x8_t a, __constrange(1,8) int b) // VQRSHRN.S16 d0,q0,#8
 {
     int8x8_t res64;
@@ -8843,7 +8845,7 @@ _NEON2SSE_INLINE int8x8_t vqrshrn_n_s16(int16x8_t a, __constrange(1,8) int b) //
     return64(r16);
 }
 
-int16x4_t vqrshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VQRSHRN.S32 d0,q0,#16
+_NEON2SSESTORAGE int16x4_t vqrshrn_n_s32(int32x4_t a, __constrange(1,16) int b); // VQRSHRN.S32 d0,q0,#16
 _NEON2SSE_INLINE int16x4_t vqrshrn_n_s32(int32x4_t a, __constrange(1,16) int b) // VQRSHRN.S32 d0,q0,#16
 {
     int16x4_t res64;
@@ -8853,7 +8855,7 @@ _NEON2SSE_INLINE int16x4_t vqrshrn_n_s32(int32x4_t a, __constrange(1,16) int b)
     return64(r32);
 }
 
-int32x2_t vqrshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VQRSHRN.S64 d0,q0,#32
+_NEON2SSESTORAGE int32x2_t vqrshrn_n_s64(int64x2_t a, __constrange(1,32) int b); // VQRSHRN.S64 d0,q0,#32
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqrshrn_n_s64(int64x2_t a, __constrange(1,32) int b), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     //no optimal SIMD solution found
@@ -8873,7 +8875,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqrshrn_n_s64(int64x2_t
     return res;
 }
 
-uint8x8_t vqrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VQRSHRN.s16 d0,q0,#8
+_NEON2SSESTORAGE uint8x8_t vqrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b); // VQRSHRN.s16 d0,q0,#8
 _NEON2SSE_INLINE uint8x8_t vqrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b) // VQRSHRN.s16 d0,q0,#8
 {
     uint8x8_t res64;
@@ -8883,7 +8885,7 @@ _NEON2SSE_INLINE uint8x8_t vqrshrn_n_u16(uint16x8_t a, __constrange(1,8) int b)
     return64(r16);
 }
 
-uint16x4_t vqrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VQRSHRN.U32 d0,q0,#16
+_NEON2SSESTORAGE uint16x4_t vqrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b); // VQRSHRN.U32 d0,q0,#16
 _NEON2SSE_INLINE uint16x4_t vqrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b) // VQRSHRN.U32 d0,q0,#16
 {
     uint16x4_t res64;
@@ -8893,7 +8895,7 @@ _NEON2SSE_INLINE uint16x4_t vqrshrn_n_u32(uint32x4_t a, __constrange(1,16) int b
     return64(r32);
 }
 
-uint32x2_t vqrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VQRSHRN.U64 d0,q0,#32
+_NEON2SSESTORAGE uint32x2_t vqrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b); // VQRSHRN.U64 d0,q0,#32
 _NEON2SSE_INLINE uint32x2_t vqrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b)
 {
     //serial solution may be faster
@@ -8910,7 +8912,7 @@ _NEON2SSE_INLINE uint32x2_t vqrshrn_n_u64(uint64x2_t a, __constrange(1,32) int b
 
 //************** Vector widening shift left by constant ****************
 //************************************************************************
-int16x8_t vshll_n_s8(int8x8_t a, __constrange(0,8) int b); // VSHLL.S8 q0,d0,#0
+_NEON2SSESTORAGE int16x8_t vshll_n_s8(int8x8_t a, __constrange(0,8) int b); // VSHLL.S8 q0,d0,#0
 _NEON2SSE_INLINE int16x8_t vshll_n_s8(int8x8_t a, __constrange(0,8) int b) // VSHLL.S8 q0,d0,#0
 {
     __m128i r;
@@ -8918,7 +8920,7 @@ _NEON2SSE_INLINE int16x8_t vshll_n_s8(int8x8_t a, __constrange(0,8) int b) // VS
     return _mm_slli_epi16 (r, b);
 }
 
-int32x4_t vshll_n_s16(int16x4_t a, __constrange(0,16) int b); // VSHLL.S16 q0,d0,#0
+_NEON2SSESTORAGE int32x4_t vshll_n_s16(int16x4_t a, __constrange(0,16) int b); // VSHLL.S16 q0,d0,#0
 _NEON2SSE_INLINE int32x4_t vshll_n_s16(int16x4_t a, __constrange(0,16) int b) // VSHLL.S16 q0,d0,#0
 {
     __m128i r;
@@ -8926,7 +8928,7 @@ _NEON2SSE_INLINE int32x4_t vshll_n_s16(int16x4_t a, __constrange(0,16) int b) //
     return _mm_slli_epi32 (r, b);
 }
 
-int64x2_t vshll_n_s32(int32x2_t a, __constrange(0,32) int b); // VSHLL.S32 q0,d0,#0
+_NEON2SSESTORAGE int64x2_t vshll_n_s32(int32x2_t a, __constrange(0,32) int b); // VSHLL.S32 q0,d0,#0
 _NEON2SSE_INLINE int64x2_t vshll_n_s32(int32x2_t a, __constrange(0,32) int b) // VSHLL.S32 q0,d0,#0
 {
     __m128i r;
@@ -8934,7 +8936,7 @@ _NEON2SSE_INLINE int64x2_t vshll_n_s32(int32x2_t a, __constrange(0,32) int b) //
     return _mm_slli_epi64 (r, b);
 }
 
-uint16x8_t vshll_n_u8(uint8x8_t a, __constrange(0,8) int b); // VSHLL.U8 q0,d0,#0
+_NEON2SSESTORAGE uint16x8_t vshll_n_u8(uint8x8_t a, __constrange(0,8) int b); // VSHLL.U8 q0,d0,#0
 _NEON2SSE_INLINE uint16x8_t vshll_n_u8(uint8x8_t a, __constrange(0,8) int b) // VSHLL.U8 q0,d0,#0
 {
     //no uint8 to uint16 conversion available, manual conversion used
@@ -8944,7 +8946,7 @@ _NEON2SSE_INLINE uint16x8_t vshll_n_u8(uint8x8_t a, __constrange(0,8) int b) //
     return _mm_slli_epi16 (r, b);
 }
 
-uint32x4_t vshll_n_u16(uint16x4_t a, __constrange(0,16) int b); // VSHLL.s16 q0,d0,#0
+_NEON2SSESTORAGE uint32x4_t vshll_n_u16(uint16x4_t a, __constrange(0,16) int b); // VSHLL.s16 q0,d0,#0
 _NEON2SSE_INLINE uint32x4_t vshll_n_u16(uint16x4_t a, __constrange(0,16) int b) // VSHLL.s16 q0,d0,#0
 {
     //no uint16 to uint32 conversion available, manual conversion used
@@ -8954,7 +8956,7 @@ _NEON2SSE_INLINE uint32x4_t vshll_n_u16(uint16x4_t a, __constrange(0,16) int b)
     return _mm_slli_epi32 (r, b);
 }
 
-uint64x2_t vshll_n_u32(uint32x2_t a, __constrange(0,32) int b); // VSHLL.U32 q0,d0,#0
+_NEON2SSESTORAGE uint64x2_t vshll_n_u32(uint32x2_t a, __constrange(0,32) int b); // VSHLL.U32 q0,d0,#0
 _NEON2SSE_INLINE uint64x2_t vshll_n_u32(uint32x2_t a, __constrange(0,32) int b) // VSHLL.U32 q0,d0,#0
 {
     //no uint32 to uint64 conversion available, manual conversion used
@@ -8973,7 +8975,7 @@ _NEON2SSE_INLINE uint64x2_t vshll_n_u32(uint32x2_t a, __constrange(0,32) int b)
 //**************** Vector shift right and insert ************************************
 //Actually the "c" left bits from "a" are the only bits remained from "a"  after the shift.
 //All other bits are taken from b shifted.
-int8x8_t vsri_n_s8(int8x8_t a,  int8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
+_NEON2SSESTORAGE int8x8_t vsri_n_s8(int8x8_t a,  int8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
 _NEON2SSE_INLINE int8x8_t vsri_n_s8(int8x8_t a,  int8x8_t b, __constrange(1,8) int c)
 {
     int8x8_t res64;
@@ -8981,7 +8983,7 @@ _NEON2SSE_INLINE int8x8_t vsri_n_s8(int8x8_t a,  int8x8_t b, __constrange(1,8) i
 }
 
 
-int16x4_t vsri_n_s16(int16x4_t a,  int16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
+_NEON2SSESTORAGE int16x4_t vsri_n_s16(int16x4_t a,  int16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
 _NEON2SSE_INLINE int16x4_t vsri_n_s16(int16x4_t a,  int16x4_t b, __constrange(1,16) int c)
 {
     int16x4_t res64;
@@ -8989,7 +8991,7 @@ _NEON2SSE_INLINE int16x4_t vsri_n_s16(int16x4_t a,  int16x4_t b, __constrange(1,
 }
 
 
-int32x2_t vsri_n_s32(int32x2_t a,  int32x2_t b, __constrange(1,32) int c); // VSRI.32 d0,d0,#32
+_NEON2SSESTORAGE int32x2_t vsri_n_s32(int32x2_t a,  int32x2_t b, __constrange(1,32) int c); // VSRI.32 d0,d0,#32
 _NEON2SSE_INLINE int32x2_t vsri_n_s32(int32x2_t a,  int32x2_t b, __constrange(1,32) int c)
 {
     int32x2_t res64;
@@ -8997,7 +8999,7 @@ _NEON2SSE_INLINE int32x2_t vsri_n_s32(int32x2_t a,  int32x2_t b, __constrange(1,
 }
 
 
-int64x1_t vsri_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VSRI.64 d0,d0,#64
+_NEON2SSESTORAGE int64x1_t vsri_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c); // VSRI.64 d0,d0,#64
 _NEON2SSE_INLINE int64x1_t vsri_n_s64(int64x1_t a, int64x1_t b, __constrange(1,64) int c)
 {
     int64x1_t res;
@@ -9009,26 +9011,26 @@ _NEON2SSE_INLINE int64x1_t vsri_n_s64(int64x1_t a, int64x1_t b, __constrange(1,6
     return res;
 }
 
-uint8x8_t vsri_n_u8(uint8x8_t a,  uint8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
+_NEON2SSESTORAGE uint8x8_t vsri_n_u8(uint8x8_t a,  uint8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
 #define vsri_n_u8 vsri_n_s8
 
-uint16x4_t vsri_n_u16(uint16x4_t a,  uint16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
+_NEON2SSESTORAGE uint16x4_t vsri_n_u16(uint16x4_t a,  uint16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
 #define vsri_n_u16 vsri_n_s16
 
-uint32x2_t vsri_n_u32(uint32x2_t a,  uint32x2_t b, __constrange(1,32) int c); // VSRI.32 d0,d0,#32
+_NEON2SSESTORAGE uint32x2_t vsri_n_u32(uint32x2_t a,  uint32x2_t b, __constrange(1,32) int c); // VSRI.32 d0,d0,#32
 #define vsri_n_u32 vsri_n_s32
 
 
-uint64x1_t vsri_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VSRI.64 d0,d0,#64
+_NEON2SSESTORAGE uint64x1_t vsri_n_u64(uint64x1_t a, uint64x1_t b, __constrange(1,64) int c); // VSRI.64 d0,d0,#64
 #define vsri_n_u64 vsri_n_s64
 
-poly8x8_t vsri_n_p8(poly8x8_t a, poly8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
+_NEON2SSESTORAGE poly8x8_t vsri_n_p8(poly8x8_t a, poly8x8_t b, __constrange(1,8) int c); // VSRI.8 d0,d0,#8
 #define vsri_n_p8 vsri_n_u8
 
-poly16x4_t vsri_n_p16(poly16x4_t a, poly16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
+_NEON2SSESTORAGE poly16x4_t vsri_n_p16(poly16x4_t a, poly16x4_t b, __constrange(1,16) int c); // VSRI.16 d0,d0,#16
 #define vsri_n_p16 vsri_n_u16
 
-int8x16_t vsriq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
+_NEON2SSESTORAGE int8x16_t vsriq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
 _NEON2SSE_INLINE int8x16_t vsriq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8) int c) // VSRI.8 q0,q0,#8
 {
     __m128i maskA, a_masked;
@@ -9040,7 +9042,7 @@ _NEON2SSE_INLINE int8x16_t vsriq_n_s8(int8x16_t a, int8x16_t b, __constrange(1,8
     return _mm_or_si128 (a_masked, b_shift); //combine (insert b into a)
 }
 
-int16x8_t vsriq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
+_NEON2SSESTORAGE int16x8_t vsriq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
 _NEON2SSE_INLINE int16x8_t vsriq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,16) int c) // VSRI.16 q0,q0,#16
 {
     //to cut "c" left bits from a we do shift right and then  shift back left providing c right zeros in a
@@ -9052,7 +9054,7 @@ _NEON2SSE_INLINE int16x8_t vsriq_n_s16(int16x8_t a, int16x8_t b, __constrange(1,
     return _mm_or_si128 (a_c, b_shift); //combine (insert b into a)
 }
 
-int32x4_t vsriq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VSRI.32 q0,q0,#32
+_NEON2SSESTORAGE int32x4_t vsriq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c); // VSRI.32 q0,q0,#32
 _NEON2SSE_INLINE int32x4_t vsriq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,32) int c) // VSRI.32 q0,q0,#32
 {
     //to cut "c" left bits from a we do shift right and then  shift back left providing c right zeros in a
@@ -9064,7 +9066,7 @@ _NEON2SSE_INLINE int32x4_t vsriq_n_s32(int32x4_t a, int32x4_t b, __constrange(1,
     return _mm_or_si128 (a_c, b_shift); //combine (insert b into a)
 }
 
-int64x2_t vsriq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VSRI.64 q0,q0,#64
+_NEON2SSESTORAGE int64x2_t vsriq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c); // VSRI.64 q0,q0,#64
 _NEON2SSE_INLINE int64x2_t vsriq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,64) int c)
 {
     //serial solution may be faster
@@ -9076,29 +9078,29 @@ _NEON2SSE_INLINE int64x2_t vsriq_n_s64(int64x2_t a, int64x2_t b, __constrange(1,
     return _mm_or_si128 (a_c, b_shift); //combine (insert b into a)
 }
 
-uint8x16_t vsriq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
+_NEON2SSESTORAGE uint8x16_t vsriq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
 #define vsriq_n_u8 vsriq_n_s8
 
-uint16x8_t vsriq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
+_NEON2SSESTORAGE uint16x8_t vsriq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
 #define vsriq_n_u16 vsriq_n_s16
 
-uint32x4_t vsriq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VSRI.32 q0,q0,#32
+_NEON2SSESTORAGE uint32x4_t vsriq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(1,32) int c); // VSRI.32 q0,q0,#32
 #define vsriq_n_u32 vsriq_n_s32
 
-uint64x2_t vsriq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VSRI.64 q0,q0,#64
+_NEON2SSESTORAGE uint64x2_t vsriq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(1,64) int c); // VSRI.64 q0,q0,#64
 #define vsriq_n_u64 vsriq_n_s64
 
-poly8x16_t vsriq_n_p8(poly8x16_t a, poly8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
+_NEON2SSESTORAGE poly8x16_t vsriq_n_p8(poly8x16_t a, poly8x16_t b, __constrange(1,8) int c); // VSRI.8 q0,q0,#8
 #define vsriq_n_p8 vsriq_n_u8
 
-poly16x8_t vsriq_n_p16(poly16x8_t a, poly16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
+_NEON2SSESTORAGE poly16x8_t vsriq_n_p16(poly16x8_t a, poly16x8_t b, __constrange(1,16) int c); // VSRI.16 q0,q0,#16
 #define vsriq_n_p16 vsriq_n_u16
 
 //***** Vector shift left and insert *********************************************
 //*********************************************************************************
 //Actually the "c" right bits from "a" are the only bits remained from "a"  after the shift.
 //All other bits are taken from b shifted. Ending zeros are inserted in b in the shift proces. We need to combine "a" and "b shifted".
-int8x8_t vsli_n_s8(int8x8_t a,  int8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
+_NEON2SSESTORAGE int8x8_t vsli_n_s8(int8x8_t a,  int8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
 _NEON2SSE_INLINE int8x8_t vsli_n_s8(int8x8_t a,  int8x8_t b, __constrange(0,7) int c)
 {
     int8x8_t res64;
@@ -9106,7 +9108,7 @@ _NEON2SSE_INLINE int8x8_t vsli_n_s8(int8x8_t a,  int8x8_t b, __constrange(0,7) i
 }
 
 
-int16x4_t vsli_n_s16(int16x4_t a,  int16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
+_NEON2SSESTORAGE int16x4_t vsli_n_s16(int16x4_t a,  int16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
 _NEON2SSE_INLINE int16x4_t vsli_n_s16(int16x4_t a,  int16x4_t b, __constrange(0,15) int c)
 {
     int16x4_t res64;
@@ -9114,14 +9116,14 @@ _NEON2SSE_INLINE int16x4_t vsli_n_s16(int16x4_t a,  int16x4_t b, __constrange(0,
 }
 
 
-int32x2_t vsli_n_s32(int32x2_t a,  int32x2_t b, __constrange(0,31) int c); // VSLI.32 d0,d0,#0
+_NEON2SSESTORAGE int32x2_t vsli_n_s32(int32x2_t a,  int32x2_t b, __constrange(0,31) int c); // VSLI.32 d0,d0,#0
 _NEON2SSE_INLINE int32x2_t vsli_n_s32(int32x2_t a,  int32x2_t b, __constrange(0,31) int c)
 {
     int32x2_t res64;
     return64(vsliq_n_s32(_pM128i(a),_pM128i(b), c));
 }
 
-int64x1_t vsli_n_s64(int64x1_t a, int64x1_t b, __constrange(0,63) int c); // VSLI.64 d0,d0,#0
+_NEON2SSESTORAGE int64x1_t vsli_n_s64(int64x1_t a, int64x1_t b, __constrange(0,63) int c); // VSLI.64 d0,d0,#0
 _NEON2SSE_INLINE int64x1_t vsli_n_s64(int64x1_t a, int64x1_t b, __constrange(0,63) int c)
 {
     int64x1_t res;
@@ -9130,25 +9132,25 @@ _NEON2SSE_INLINE int64x1_t vsli_n_s64(int64x1_t a, int64x1_t b, __constrange(0,6
 }
 
 
-uint8x8_t vsli_n_u8(uint8x8_t a,  uint8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
+_NEON2SSESTORAGE uint8x8_t vsli_n_u8(uint8x8_t a,  uint8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
 #define vsli_n_u8 vsli_n_s8
 
-uint16x4_t vsli_n_u16(uint16x4_t a,  uint16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
+_NEON2SSESTORAGE uint16x4_t vsli_n_u16(uint16x4_t a,  uint16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
 #define vsli_n_u16 vsli_n_s16
 
-uint32x2_t vsli_n_u32(uint32x2_t a,  uint32x2_t b, __constrange(0,31) int c); // VSLI.32 d0,d0,#0
+_NEON2SSESTORAGE uint32x2_t vsli_n_u32(uint32x2_t a,  uint32x2_t b, __constrange(0,31) int c); // VSLI.32 d0,d0,#0
 #define vsli_n_u32 vsli_n_s32
 
-uint64x1_t vsli_n_u64(uint64x1_t a, uint64x1_t b, __constrange(0,63) int c); // VSLI.64 d0,d0,#0
+_NEON2SSESTORAGE uint64x1_t vsli_n_u64(uint64x1_t a, uint64x1_t b, __constrange(0,63) int c); // VSLI.64 d0,d0,#0
 #define vsli_n_u64 vsli_n_s64
 
-poly8x8_t vsli_n_p8(poly8x8_t a, poly8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
+_NEON2SSESTORAGE poly8x8_t vsli_n_p8(poly8x8_t a, poly8x8_t b, __constrange(0,7) int c); // VSLI.8 d0,d0,#0
 #define vsli_n_p8 vsli_n_u8
 
-poly16x4_t vsli_n_p16(poly16x4_t a, poly16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
+_NEON2SSESTORAGE poly16x4_t vsli_n_p16(poly16x4_t a, poly16x4_t b, __constrange(0,15) int c); // VSLI.16 d0,d0,#0
 #define vsli_n_p16 vsli_n_u16
 
-int8x16_t vsliq_n_s8(int8x16_t a, int8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
+_NEON2SSESTORAGE int8x16_t vsliq_n_s8(int8x16_t a, int8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
 _NEON2SSE_INLINE int8x16_t vsliq_n_s8(int8x16_t a, int8x16_t b, __constrange(0,7) int c) // VSLI.8 q0,q0,#0
 {
     __m128i maskA, a_masked;
@@ -9160,7 +9162,7 @@ _NEON2SSE_INLINE int8x16_t vsliq_n_s8(int8x16_t a, int8x16_t b, __constrange(0,7
     return _mm_or_si128 (b_shift, a_masked); //combine (insert b into a)
 }
 
-int16x8_t vsliq_n_s16(int16x8_t a, int16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
+_NEON2SSESTORAGE int16x8_t vsliq_n_s16(int16x8_t a, int16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
 _NEON2SSE_INLINE int16x8_t vsliq_n_s16(int16x8_t a, int16x8_t b, __constrange(0,15) int c) // VSLI.16 q0,q0,#0
 {
     //to cut "c" right bits from a we do shift left and then logical shift back right providing (16-c)zeros in a
@@ -9172,7 +9174,7 @@ _NEON2SSE_INLINE int16x8_t vsliq_n_s16(int16x8_t a, int16x8_t b, __constrange(0,
     return _mm_or_si128 (b_shift, a_c); //combine (insert b into a)
 }
 
-int32x4_t vsliq_n_s32(int32x4_t a, int32x4_t b, __constrange(0,31) int c); // VSLI.32 q0,q0,#0
+_NEON2SSESTORAGE int32x4_t vsliq_n_s32(int32x4_t a, int32x4_t b, __constrange(0,31) int c); // VSLI.32 q0,q0,#0
 _NEON2SSE_INLINE int32x4_t vsliq_n_s32(int32x4_t a, int32x4_t b, __constrange(0,31) int c) // VSLI.32 q0,q0,#0
 {
     //solution may be  not optimal compared with the serial one
@@ -9185,7 +9187,7 @@ _NEON2SSE_INLINE int32x4_t vsliq_n_s32(int32x4_t a, int32x4_t b, __constrange(0,
     return _mm_or_si128 (b_shift, a_c); //combine (insert b into a)
 }
 
-int64x2_t vsliq_n_s64(int64x2_t a, int64x2_t b, __constrange(0,63) int c); // VSLI.64 q0,q0,#0
+_NEON2SSESTORAGE int64x2_t vsliq_n_s64(int64x2_t a, int64x2_t b, __constrange(0,63) int c); // VSLI.64 q0,q0,#0
 _NEON2SSE_INLINE int64x2_t vsliq_n_s64(int64x2_t a, int64x2_t b, __constrange(0,63) int c) // VSLI.64 q0,q0,#0
 {
     //solution may be  not optimal compared with the serial one
@@ -9198,22 +9200,22 @@ _NEON2SSE_INLINE int64x2_t vsliq_n_s64(int64x2_t a, int64x2_t b, __constrange(0,
     return _mm_or_si128 (b_shift, a_c); //combine (insert b into a)
 }
 
-uint8x16_t vsliq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
+_NEON2SSESTORAGE uint8x16_t vsliq_n_u8(uint8x16_t a, uint8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
 #define vsliq_n_u8 vsliq_n_s8
 
-uint16x8_t vsliq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
+_NEON2SSESTORAGE uint16x8_t vsliq_n_u16(uint16x8_t a, uint16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
 #define vsliq_n_u16 vsliq_n_s16
 
-uint32x4_t vsliq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(0,31) int c); // VSLI.32 q0,q0,#0
+_NEON2SSESTORAGE uint32x4_t vsliq_n_u32(uint32x4_t a, uint32x4_t b, __constrange(0,31) int c); // VSLI.32 q0,q0,#0
 #define vsliq_n_u32 vsliq_n_s32
 
-uint64x2_t vsliq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(0,63) int c); // VSLI.64 q0,q0,#0
+_NEON2SSESTORAGE uint64x2_t vsliq_n_u64(uint64x2_t a, uint64x2_t b, __constrange(0,63) int c); // VSLI.64 q0,q0,#0
 #define vsliq_n_u64 vsliq_n_s64
 
-poly8x16_t vsliq_n_p8(poly8x16_t a, poly8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
+_NEON2SSESTORAGE poly8x16_t vsliq_n_p8(poly8x16_t a, poly8x16_t b, __constrange(0,7) int c); // VSLI.8 q0,q0,#0
 #define vsliq_n_p8 vsliq_n_u8
 
-poly16x8_t vsliq_n_p16(poly16x8_t a, poly16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
+_NEON2SSESTORAGE poly16x8_t vsliq_n_p16(poly16x8_t a, poly16x8_t b, __constrange(0,15) int c); // VSLI.16 q0,q0,#0
 #define vsliq_n_p16 vsliq_n_u16
 
 // ***********************************************************************************************
@@ -9229,31 +9231,31 @@ poly16x8_t vsliq_n_p16(poly16x8_t a, poly16x8_t b, __constrange(0,15) int c); //
 #define LOAD_SI128(ptr) \
         ( ((uintptr_t)(ptr) & 15) == 0 ) ? _mm_load_si128((__m128i*)(ptr)) : _mm_loadu_si128((__m128i*)(ptr))
 
-uint8x16_t vld1q_u8(__transfersize(16) uint8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint8x16_t vld1q_u8(__transfersize(16) uint8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
 #define vld1q_u8 LOAD_SI128
 
-uint16x8_t vld1q_u16(__transfersize(8) uint16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint16x8_t vld1q_u16(__transfersize(8) uint16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
 #define vld1q_u16 LOAD_SI128
 
-uint32x4_t vld1q_u32(__transfersize(4) uint32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint32x4_t vld1q_u32(__transfersize(4) uint32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
 #define vld1q_u32 LOAD_SI128
 
-uint64x2_t vld1q_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint64x2_t vld1q_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
 #define vld1q_u64 LOAD_SI128
 
-int8x16_t vld1q_s8(__transfersize(16) int8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE int8x16_t vld1q_s8(__transfersize(16) int8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
 #define vld1q_s8 LOAD_SI128
 
-int16x8_t vld1q_s16(__transfersize(8) int16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE int16x8_t vld1q_s16(__transfersize(8) int16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
 #define vld1q_s16 LOAD_SI128
 
-int32x4_t vld1q_s32(__transfersize(4) int32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE int32x4_t vld1q_s32(__transfersize(4) int32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
 #define vld1q_s32 LOAD_SI128
 
-int64x2_t vld1q_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE int64x2_t vld1q_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
 #define vld1q_s64 LOAD_SI128
 
-float16x8_t vld1q_f16(__transfersize(8) __fp16 const * ptr); // VLD1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE float16x8_t vld1q_f16(__transfersize(8) __fp16 const * ptr); // VLD1.16 {d0, d1}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers
 /* _NEON2SSE_INLINE float16x8_t vld1q_f16(__transfersize(8) __fp16 const * ptr)// VLD1.16 {d0, d1}, [r0]
 {__m128 f1 = _mm_set_ps (ptr[3], ptr[2], ptr[1], ptr[0]);
@@ -9261,7 +9263,7 @@ __m128 f2;
 f2 = _mm_set_ps (ptr[7], ptr[6], ptr[5], ptr[4]);
 }*/
 
-float32x4_t vld1q_f32(__transfersize(4) float32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE float32x4_t vld1q_f32(__transfersize(4) float32_t const * ptr); // VLD1.32 {d0, d1}, [r0]
 _NEON2SSE_INLINE float32x4_t vld1q_f32(__transfersize(4) float32_t const * ptr)
 {
     if( (((uintptr_t)(ptr)) & 15 ) == 0 ) //16 bits aligned
@@ -9270,41 +9272,41 @@ _NEON2SSE_INLINE float32x4_t vld1q_f32(__transfersize(4) float32_t const * ptr)
         return _mm_loadu_ps(ptr);
 }
 
-poly8x16_t vld1q_p8(__transfersize(16) poly8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE poly8x16_t vld1q_p8(__transfersize(16) poly8_t const * ptr); // VLD1.8 {d0, d1}, [r0]
 #define vld1q_p8  LOAD_SI128
 
-poly16x8_t vld1q_p16(__transfersize(8) poly16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE poly16x8_t vld1q_p16(__transfersize(8) poly16_t const * ptr); // VLD1.16 {d0, d1}, [r0]
 #define vld1q_p16 LOAD_SI128
 
-uint8x8_t vld1_u8(__transfersize(8) uint8_t const * ptr); // VLD1.8 {d0}, [r0]
+_NEON2SSESTORAGE uint8x8_t vld1_u8(__transfersize(8) uint8_t const * ptr); // VLD1.8 {d0}, [r0]
 #define vld1_u8(ptr)  *((__m64_128*)(ptr)) //was _mm_loadl_epi64((__m128i*)(ptr))
 
-uint16x4_t vld1_u16(__transfersize(4) uint16_t const * ptr); // VLD1.16 {d0}, [r0]
+_NEON2SSESTORAGE uint16x4_t vld1_u16(__transfersize(4) uint16_t const * ptr); // VLD1.16 {d0}, [r0]
 #define vld1_u16 vld1_u8
 
-uint32x2_t vld1_u32(__transfersize(2) uint32_t const * ptr); // VLD1.32 {d0}, [r0]
+_NEON2SSESTORAGE uint32x2_t vld1_u32(__transfersize(2) uint32_t const * ptr); // VLD1.32 {d0}, [r0]
 #define vld1_u32 vld1_u8
 
 
-uint64x1_t vld1_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE uint64x1_t vld1_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
 #define vld1_u64 vld1_u8
 
-int8x8_t vld1_s8(__transfersize(8) int8_t const * ptr); // VLD1.8 {d0}, [r0]
+_NEON2SSESTORAGE int8x8_t vld1_s8(__transfersize(8) int8_t const * ptr); // VLD1.8 {d0}, [r0]
 #define vld1_s8 vld1_u8
 
-int16x4_t vld1_s16(__transfersize(4) int16_t const * ptr); // VLD1.16 {d0}, [r0]
+_NEON2SSESTORAGE int16x4_t vld1_s16(__transfersize(4) int16_t const * ptr); // VLD1.16 {d0}, [r0]
 #define vld1_s16 vld1_u16
 
-int32x2_t vld1_s32(__transfersize(2) int32_t const * ptr); // VLD1.32 {d0}, [r0]
+_NEON2SSESTORAGE int32x2_t vld1_s32(__transfersize(2) int32_t const * ptr); // VLD1.32 {d0}, [r0]
 #define vld1_s32 vld1_u32
 
-int64x1_t vld1_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE int64x1_t vld1_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
 #define vld1_s64 vld1_u64
 
-float16x4_t vld1_f16(__transfersize(4) __fp16 const * ptr); // VLD1.16 {d0}, [r0]
+_NEON2SSESTORAGE float16x4_t vld1_f16(__transfersize(4) __fp16 const * ptr); // VLD1.16 {d0}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit like _mm_set_ps (ptr[3], ptr[2], ptr[1], ptr[0]);
 
-float32x2_t vld1_f32(__transfersize(2) float32_t const * ptr); // VLD1.32 {d0}, [r0]
+_NEON2SSESTORAGE float32x2_t vld1_f32(__transfersize(2) float32_t const * ptr); // VLD1.32 {d0}, [r0]
 _NEON2SSE_INLINE float32x2_t vld1_f32(__transfersize(2) float32_t const * ptr)
 {
     float32x2_t res;
@@ -9313,14 +9315,14 @@ _NEON2SSE_INLINE float32x2_t vld1_f32(__transfersize(2) float32_t const * ptr)
     return res;
 }
 
-poly8x8_t vld1_p8(__transfersize(8) poly8_t const * ptr); // VLD1.8 {d0}, [r0]
+_NEON2SSESTORAGE poly8x8_t vld1_p8(__transfersize(8) poly8_t const * ptr); // VLD1.8 {d0}, [r0]
 #define vld1_p8 vld1_u8
 
-poly16x4_t vld1_p16(__transfersize(4) poly16_t const * ptr); // VLD1.16 {d0}, [r0]
+_NEON2SSESTORAGE poly16x4_t vld1_p16(__transfersize(4) poly16_t const * ptr); // VLD1.16 {d0}, [r0]
 #define vld1_p16 vld1_u16
 
 
-float64x2_t vld1q_f64(__transfersize(4) float64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE float64x2_t vld1q_f64(__transfersize(4) float64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
 _NEON2SSE_INLINE float64x2_t vld1q_f64(__transfersize(4) float64_t const * ptr)
 {
     if ((((uintptr_t)(ptr)) & 15) == 0) //16 bits aligned
@@ -9333,32 +9335,32 @@ _NEON2SSE_INLINE float64x2_t vld1q_f64(__transfersize(4) float64_t const * ptr)
 //***********************************************************************************************************
 //******* Lane load functions - insert the data at  vector's given position (lane) *************************
 //***********************************************************************************************************
-uint8x16_t vld1q_lane_u8(__transfersize(1) uint8_t const * ptr, uint8x16_t vec, __constrange(0,15) int lane); // VLD1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint8x16_t vld1q_lane_u8(__transfersize(1) uint8_t const * ptr, uint8x16_t vec, __constrange(0,15) int lane); // VLD1.8 {d0[0]}, [r0]
 #define vld1q_lane_u8(ptr, vec, lane) _MM_INSERT_EPI8(vec, *(ptr), lane)
 
-uint16x8_t vld1q_lane_u16(__transfersize(1)    uint16_t const * ptr, uint16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint16x8_t vld1q_lane_u16(__transfersize(1)    uint16_t const * ptr, uint16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
 #define vld1q_lane_u16(ptr, vec, lane) _MM_INSERT_EPI16(vec, *(ptr), lane)
 
-uint32x4_t vld1q_lane_u32(__transfersize(1) uint32_t const * ptr, uint32x4_t vec, __constrange(0,3) int lane); // VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint32x4_t vld1q_lane_u32(__transfersize(1) uint32_t const * ptr, uint32x4_t vec, __constrange(0,3) int lane); // VLD1.32 {d0[0]}, [r0]
 #define vld1q_lane_u32(ptr, vec, lane) _MM_INSERT_EPI32(vec, *(ptr), lane)
 
-uint64x2_t vld1q_lane_u64(__transfersize(1) uint64_t const * ptr, uint64x2_t vec, __constrange(0,1) int lane); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE uint64x2_t vld1q_lane_u64(__transfersize(1) uint64_t const * ptr, uint64x2_t vec, __constrange(0,1) int lane); // VLD1.64 {d0}, [r0]
 #define vld1q_lane_u64(ptr, vec, lane) _MM_INSERT_EPI64(vec, *(ptr), lane); // _p;
 
 
-int8x16_t vld1q_lane_s8(__transfersize(1) int8_t const * ptr, int8x16_t vec, __constrange(0,15) int lane); // VLD1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE int8x16_t vld1q_lane_s8(__transfersize(1) int8_t const * ptr, int8x16_t vec, __constrange(0,15) int lane); // VLD1.8 {d0[0]}, [r0]
 #define vld1q_lane_s8(ptr, vec, lane) _MM_INSERT_EPI8(vec, *(ptr), lane)
 
-int16x8_t vld1q_lane_s16(__transfersize(1) int16_t const * ptr, int16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE int16x8_t vld1q_lane_s16(__transfersize(1) int16_t const * ptr, int16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
 #define vld1q_lane_s16(ptr, vec, lane) _MM_INSERT_EPI16(vec, *(ptr), lane)
 
-int32x4_t vld1q_lane_s32(__transfersize(1) int32_t const * ptr, int32x4_t vec, __constrange(0,3) int lane); // VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE int32x4_t vld1q_lane_s32(__transfersize(1) int32_t const * ptr, int32x4_t vec, __constrange(0,3) int lane); // VLD1.32 {d0[0]}, [r0]
 #define vld1q_lane_s32(ptr, vec, lane) _MM_INSERT_EPI32(vec, *(ptr), lane)
 
-float16x8_t vld1q_lane_f16(__transfersize(1) __fp16 const * ptr, float16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE float16x8_t vld1q_lane_f16(__transfersize(1) __fp16 const * ptr, float16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
 //current IA SIMD doesn't support float16
 
-float32x4_t vld1q_lane_f32(__transfersize(1) float32_t const * ptr, float32x4_t vec, __constrange(0,3) int lane); // VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE float32x4_t vld1q_lane_f32(__transfersize(1) float32_t const * ptr, float32x4_t vec, __constrange(0,3) int lane); // VLD1.32 {d0[0]}, [r0]
 _NEON2SSE_INLINE float32x4_t vld1q_lane_f32(__transfersize(1) float32_t const * ptr, float32x4_t vec, __constrange(0,3) int lane)
 {
     //we need to deal with  ptr  16bit NOT aligned case
@@ -9367,16 +9369,16 @@ _NEON2SSE_INLINE float32x4_t vld1q_lane_f32(__transfersize(1) float32_t const *
     return _MM_INSERT_PS(vec,  p, _INSERTPS_NDX(0, lane));
 }
 
-int64x2_t vld1q_lane_s64(__transfersize(1) int64_t const * ptr, int64x2_t vec, __constrange(0,1) int lane); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE int64x2_t vld1q_lane_s64(__transfersize(1) int64_t const * ptr, int64x2_t vec, __constrange(0,1) int lane); // VLD1.64 {d0}, [r0]
 #define vld1q_lane_s64(ptr, vec, lane) _MM_INSERT_EPI64(vec, *(ptr), lane)
 
-poly8x16_t vld1q_lane_p8(__transfersize(1) poly8_t const * ptr, poly8x16_t vec, __constrange(0,15) int lane); // VLD1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE poly8x16_t vld1q_lane_p8(__transfersize(1) poly8_t const * ptr, poly8x16_t vec, __constrange(0,15) int lane); // VLD1.8 {d0[0]}, [r0]
 #define vld1q_lane_p8(ptr, vec, lane) _MM_INSERT_EPI8(vec, *(ptr), lane)
 
-poly16x8_t vld1q_lane_p16(__transfersize(1) poly16_t const * ptr, poly16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE poly16x8_t vld1q_lane_p16(__transfersize(1) poly16_t const * ptr, poly16x8_t vec, __constrange(0,7) int lane); // VLD1.16 {d0[0]}, [r0]
 #define vld1q_lane_p16(ptr, vec, lane) _MM_INSERT_EPI16(vec, *(ptr), lane)
 
-uint8x8_t vld1_lane_u8(__transfersize(1) uint8_t const * ptr, uint8x8_t vec, __constrange(0,7) int lane); // VLD1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint8x8_t vld1_lane_u8(__transfersize(1) uint8_t const * ptr, uint8x8_t vec, __constrange(0,7) int lane); // VLD1.8 {d0[0]}, [r0]
 _NEON2SSE_INLINE uint8x8_t vld1_lane_u8(__transfersize(1) uint8_t const * ptr, uint8x8_t vec, __constrange(0,7) int lane)
 {
     uint8x8_t res;
@@ -9385,7 +9387,7 @@ _NEON2SSE_INLINE uint8x8_t vld1_lane_u8(__transfersize(1) uint8_t const * ptr, u
     return res;
 }
 
-uint16x4_t vld1_lane_u16(__transfersize(1) uint16_t const * ptr, uint16x4_t vec, __constrange(0,3) int lane); // VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint16x4_t vld1_lane_u16(__transfersize(1) uint16_t const * ptr, uint16x4_t vec, __constrange(0,3) int lane); // VLD1.16 {d0[0]}, [r0]
 _NEON2SSE_INLINE uint16x4_t vld1_lane_u16(__transfersize(1) uint16_t const * ptr, uint16x4_t vec, __constrange(0,3) int lane)
 {
     uint16x4_t res;
@@ -9394,7 +9396,7 @@ _NEON2SSE_INLINE uint16x4_t vld1_lane_u16(__transfersize(1) uint16_t const * ptr
     return res;
 }
 
-uint32x2_t vld1_lane_u32(__transfersize(1) uint32_t const * ptr, uint32x2_t vec, __constrange(0,1) int lane); // VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE uint32x2_t vld1_lane_u32(__transfersize(1) uint32_t const * ptr, uint32x2_t vec, __constrange(0,1) int lane); // VLD1.32 {d0[0]}, [r0]
 _NEON2SSE_INLINE uint32x2_t vld1_lane_u32(__transfersize(1) uint32_t const * ptr, uint32x2_t vec, __constrange(0,1) int lane)
 {
     uint32x2_t res;
@@ -9403,7 +9405,7 @@ _NEON2SSE_INLINE uint32x2_t vld1_lane_u32(__transfersize(1) uint32_t const * ptr
     return res;
 }
 
-uint64x1_t vld1_lane_u64(__transfersize(1) uint64_t const * ptr, uint64x1_t vec, __constrange(0,0) int lane); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE uint64x1_t vld1_lane_u64(__transfersize(1) uint64_t const * ptr, uint64x1_t vec, __constrange(0,0) int lane); // VLD1.64 {d0}, [r0]
 _NEON2SSE_INLINE uint64x1_t vld1_lane_u64(__transfersize(1) uint64_t const * ptr, uint64x1_t vec, __constrange(0,0) int lane)
 {
     uint64x1_t res;
@@ -9412,19 +9414,19 @@ _NEON2SSE_INLINE uint64x1_t vld1_lane_u64(__transfersize(1) uint64_t const * ptr
 }
 
 
-int8x8_t vld1_lane_s8(__transfersize(1) int8_t const * ptr, int8x8_t vec, __constrange(0,7) int lane); // VLD1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE int8x8_t vld1_lane_s8(__transfersize(1) int8_t const * ptr, int8x8_t vec, __constrange(0,7) int lane); // VLD1.8 {d0[0]}, [r0]
 #define vld1_lane_s8(ptr, vec, lane) vld1_lane_u8((uint8_t*)ptr, vec, lane)
 
-int16x4_t vld1_lane_s16(__transfersize(1) int16_t const * ptr, int16x4_t vec, __constrange(0,3) int lane); // VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE int16x4_t vld1_lane_s16(__transfersize(1) int16_t const * ptr, int16x4_t vec, __constrange(0,3) int lane); // VLD1.16 {d0[0]}, [r0]
 #define vld1_lane_s16(ptr, vec, lane) vld1_lane_u16((uint16_t*)ptr, vec, lane)
 
-int32x2_t vld1_lane_s32(__transfersize(1) int32_t const * ptr, int32x2_t vec, __constrange(0,1) int lane); // VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE int32x2_t vld1_lane_s32(__transfersize(1) int32_t const * ptr, int32x2_t vec, __constrange(0,1) int lane); // VLD1.32 {d0[0]}, [r0]
 #define vld1_lane_s32(ptr, vec, lane) vld1_lane_u32((uint32_t*)ptr, vec, lane)
 
-float16x4_t vld1_lane_f16(__transfersize(1) __fp16 const * ptr, float16x4_t vec, __constrange(0,3) int lane); // VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE float16x4_t vld1_lane_f16(__transfersize(1) __fp16 const * ptr, float16x4_t vec, __constrange(0,3) int lane); // VLD1.16 {d0[0]}, [r0]
 //current IA SIMD doesn't support float16
 
-float32x2_t vld1_lane_f32(__transfersize(1) float32_t const * ptr, float32x2_t vec, __constrange(0,1) int lane); // VLD1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE float32x2_t vld1_lane_f32(__transfersize(1) float32_t const * ptr, float32x2_t vec, __constrange(0,1) int lane); // VLD1.32 {d0[0]}, [r0]
 _NEON2SSE_INLINE float32x2_t vld1_lane_f32(__transfersize(1) float32_t const * ptr, float32x2_t vec, __constrange(0,1) int lane)
 {
     float32x2_t res;
@@ -9433,58 +9435,58 @@ _NEON2SSE_INLINE float32x2_t vld1_lane_f32(__transfersize(1) float32_t const * p
     return res;
 }
 
-int64x1_t vld1_lane_s64(__transfersize(1) int64_t const * ptr, int64x1_t vec, __constrange(0,0) int lane); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE int64x1_t vld1_lane_s64(__transfersize(1) int64_t const * ptr, int64x1_t vec, __constrange(0,0) int lane); // VLD1.64 {d0}, [r0]
 #define vld1_lane_s64(ptr, vec, lane) vld1_lane_u64((uint64_t*)ptr, vec, lane)
 
-poly8x8_t vld1_lane_p8(__transfersize(1) poly8_t const * ptr, poly8x8_t vec, __constrange(0,7) int lane); // VLD1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE poly8x8_t vld1_lane_p8(__transfersize(1) poly8_t const * ptr, poly8x8_t vec, __constrange(0,7) int lane); // VLD1.8 {d0[0]}, [r0]
 #define vld1_lane_p8 vld1_lane_u8
 
-poly16x4_t vld1_lane_p16(__transfersize(1) poly16_t const * ptr, poly16x4_t vec, __constrange(0,3) int lane); // VLD1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE poly16x4_t vld1_lane_p16(__transfersize(1) poly16_t const * ptr, poly16x4_t vec, __constrange(0,3) int lane); // VLD1.16 {d0[0]}, [r0]
 #define vld1_lane_p16 vld1_lane_s16
 
 // ****************** Load single value ( set all lanes of vector with same value from memory)**********************
 // ******************************************************************************************************************
-uint8x16_t vld1q_dup_u8(__transfersize(1) uint8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE uint8x16_t vld1q_dup_u8(__transfersize(1) uint8_t const * ptr); // VLD1.8 {d0[]}, [r0]
 #define vld1q_dup_u8(ptr) _mm_set1_epi8(*(ptr))
 
-uint16x8_t vld1q_dup_u16(__transfersize(1) uint16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE uint16x8_t vld1q_dup_u16(__transfersize(1) uint16_t const * ptr); // VLD1.16 {d0[]}, [r0]
 #define vld1q_dup_u16(ptr) _mm_set1_epi16(*(ptr))
 
-uint32x4_t vld1q_dup_u32(__transfersize(1) uint32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE uint32x4_t vld1q_dup_u32(__transfersize(1) uint32_t const * ptr); // VLD1.32 {d0[]}, [r0]
 #define vld1q_dup_u32(ptr) _mm_set1_epi32(*(ptr))
 
-uint64x2_t vld1q_dup_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE uint64x2_t vld1q_dup_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
 _NEON2SSE_INLINE uint64x2_t   vld1q_dup_u64(__transfersize(1) uint64_t const * ptr)
 {
     _NEON2SSE_ALIGN_16 uint64_t val[2] = {*(ptr), *(ptr)};
     return LOAD_SI128(val);
 }
 
-int8x16_t vld1q_dup_s8(__transfersize(1) int8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE int8x16_t vld1q_dup_s8(__transfersize(1) int8_t const * ptr); // VLD1.8 {d0[]}, [r0]
 #define vld1q_dup_s8(ptr) _mm_set1_epi8(*(ptr))
 
-int16x8_t vld1q_dup_s16(__transfersize(1) int16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE int16x8_t vld1q_dup_s16(__transfersize(1) int16_t const * ptr); // VLD1.16 {d0[]}, [r0]
 #define vld1q_dup_s16(ptr) _mm_set1_epi16 (*(ptr))
 
-int32x4_t vld1q_dup_s32(__transfersize(1) int32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE int32x4_t vld1q_dup_s32(__transfersize(1) int32_t const * ptr); // VLD1.32 {d0[]}, [r0]
 #define vld1q_dup_s32(ptr) _mm_set1_epi32 (*(ptr))
 
-int64x2_t vld1q_dup_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE int64x2_t vld1q_dup_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
 #define vld1q_dup_s64(ptr) vld1q_dup_u64((uint64_t*)ptr)
 
-float16x8_t vld1q_dup_f16(__transfersize(1) __fp16 const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE float16x8_t vld1q_dup_f16(__transfersize(1) __fp16 const * ptr); // VLD1.16 {d0[]}, [r0]
 //current IA SIMD doesn't support float16, need to go to 32 bits
 
-float32x4_t vld1q_dup_f32(__transfersize(1) float32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE float32x4_t vld1q_dup_f32(__transfersize(1) float32_t const * ptr); // VLD1.32 {d0[]}, [r0]
 #define vld1q_dup_f32(ptr) _mm_set1_ps (*(ptr))
 
-poly8x16_t vld1q_dup_p8(__transfersize(1) poly8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE poly8x16_t vld1q_dup_p8(__transfersize(1) poly8_t const * ptr); // VLD1.8 {d0[]}, [r0]
 #define vld1q_dup_p8(ptr) _mm_set1_epi8(*(ptr))
 
-poly16x8_t vld1q_dup_p16(__transfersize(1) poly16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE poly16x8_t vld1q_dup_p16(__transfersize(1) poly16_t const * ptr); // VLD1.16 {d0[]}, [r0]
 #define vld1q_dup_p16(ptr) _mm_set1_epi16 (*(ptr))
 
-uint8x8_t vld1_dup_u8(__transfersize(1) uint8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE uint8x8_t vld1_dup_u8(__transfersize(1) uint8_t const * ptr); // VLD1.8 {d0[]}, [r0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x8_t vld1_dup_u8(__transfersize(1) uint8_t const * ptr), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     uint8x8_t res;
@@ -9495,7 +9497,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x8_t vld1_dup_u8(__transfers
     return res;
 }
 
-uint16x4_t vld1_dup_u16(__transfersize(1) uint16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE uint16x4_t vld1_dup_u16(__transfersize(1) uint16_t const * ptr); // VLD1.16 {d0[]}, [r0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x4_t vld1_dup_u16(__transfersize(1) uint16_t const * ptr), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     uint16x4_t res;
@@ -9506,7 +9508,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x4_t vld1_dup_u16(__transfe
     return res;
 }
 
-uint32x2_t vld1_dup_u32(__transfersize(1) uint32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE uint32x2_t vld1_dup_u32(__transfersize(1) uint32_t const * ptr); // VLD1.32 {d0[]}, [r0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vld1_dup_u32(__transfersize(1) uint32_t const * ptr), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     uint32x2_t res;
@@ -9515,7 +9517,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t vld1_dup_u32(__transfe
     return res;
 }
 
-uint64x1_t vld1_dup_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE uint64x1_t vld1_dup_u64(__transfersize(1) uint64_t const * ptr); // VLD1.64 {d0}, [r0]
 _NEON2SSE_INLINE uint64x1_t vld1_dup_u64(__transfersize(1) uint64_t const * ptr)
 {
     uint64x1_t res;
@@ -9523,25 +9525,25 @@ _NEON2SSE_INLINE uint64x1_t vld1_dup_u64(__transfersize(1) uint64_t const * ptr)
     return res;
 }
 
-int8x8_t vld1_dup_s8(__transfersize(1) int8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE int8x8_t vld1_dup_s8(__transfersize(1) int8_t const * ptr); // VLD1.8 {d0[]}, [r0]
 #define vld1_dup_s8(ptr) vld1_dup_u8((uint8_t*)ptr)
 
 
-int16x4_t vld1_dup_s16(__transfersize(1) int16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE int16x4_t vld1_dup_s16(__transfersize(1) int16_t const * ptr); // VLD1.16 {d0[]}, [r0]
 #define vld1_dup_s16(ptr) vld1_dup_u16((uint16_t*)ptr)
 
 
-int32x2_t vld1_dup_s32(__transfersize(1) int32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE int32x2_t vld1_dup_s32(__transfersize(1) int32_t const * ptr); // VLD1.32 {d0[]}, [r0]
 #define vld1_dup_s32(ptr) vld1_dup_u32((uint32_t*)ptr)
 
 
-int64x1_t vld1_dup_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
+_NEON2SSESTORAGE int64x1_t vld1_dup_s64(__transfersize(1) int64_t const * ptr); // VLD1.64 {d0}, [r0]
 #define vld1_dup_s64(ptr) vld1_dup_u64((uint64_t*)ptr)
 
-float16x4_t vld1_dup_f16(__transfersize(1) __fp16 const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE float16x4_t vld1_dup_f16(__transfersize(1) __fp16 const * ptr); // VLD1.16 {d0[]}, [r0]
 //current IA SIMD doesn't support float16
 
-float32x2_t vld1_dup_f32(__transfersize(1) float32_t const * ptr); // VLD1.32 {d0[]}, [r0]
+_NEON2SSESTORAGE float32x2_t vld1_dup_f32(__transfersize(1) float32_t const * ptr); // VLD1.32 {d0[]}, [r0]
 _NEON2SSE_INLINE float32x2_t vld1_dup_f32(__transfersize(1) float32_t const * ptr)
 {
     float32x2_t res;
@@ -9550,11 +9552,11 @@ _NEON2SSE_INLINE float32x2_t vld1_dup_f32(__transfersize(1) float32_t const * pt
     return res; // use last 64bits only
 }
 
-poly8x8_t vld1_dup_p8(__transfersize(1) poly8_t const * ptr); // VLD1.8 {d0[]}, [r0]
+_NEON2SSESTORAGE poly8x8_t vld1_dup_p8(__transfersize(1) poly8_t const * ptr); // VLD1.8 {d0[]}, [r0]
 #define vld1_dup_p8 vld1_dup_u8
 
 
-poly16x4_t vld1_dup_p16(__transfersize(1) poly16_t const * ptr); // VLD1.16 {d0[]}, [r0]
+_NEON2SSESTORAGE poly16x4_t vld1_dup_p16(__transfersize(1) poly16_t const * ptr); // VLD1.16 {d0[]}, [r0]
 #define vld1_dup_p16 vld1_dup_u16
 
 
@@ -9566,34 +9568,34 @@ poly16x4_t vld1_dup_p16(__transfersize(1) poly16_t const * ptr); // VLD1.16 {d0[
 #define STORE_SI128(ptr, val) \
         (((uintptr_t)(ptr) & 15) == 0 ) ? _mm_store_si128 ((__m128i*)(ptr), val) : _mm_storeu_si128 ((__m128i*)(ptr), val);
 
-void vst1q_u8(__transfersize(16) uint8_t * ptr, uint8x16_t val); // VST1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_u8(__transfersize(16) uint8_t * ptr, uint8x16_t val); // VST1.8 {d0, d1}, [r0]
 #define vst1q_u8 STORE_SI128
 
-void vst1q_u16(__transfersize(8) uint16_t * ptr, uint16x8_t val); // VST1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_u16(__transfersize(8) uint16_t * ptr, uint16x8_t val); // VST1.16 {d0, d1}, [r0]
 #define vst1q_u16 STORE_SI128
 
-void vst1q_u32(__transfersize(4) uint32_t * ptr, uint32x4_t val); // VST1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_u32(__transfersize(4) uint32_t * ptr, uint32x4_t val); // VST1.32 {d0, d1}, [r0]
 #define vst1q_u32 STORE_SI128
 
-void vst1q_u64(__transfersize(2) uint64_t * ptr, uint64x2_t val); // VST1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_u64(__transfersize(2) uint64_t * ptr, uint64x2_t val); // VST1.64 {d0, d1}, [r0]
 #define vst1q_u64 STORE_SI128
 
-void vst1q_s8(__transfersize(16) int8_t * ptr, int8x16_t val); // VST1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_s8(__transfersize(16) int8_t * ptr, int8x16_t val); // VST1.8 {d0, d1}, [r0]
 #define vst1q_s8 STORE_SI128
 
-void vst1q_s16(__transfersize(8) int16_t * ptr, int16x8_t val); // VST1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_s16(__transfersize(8) int16_t * ptr, int16x8_t val); // VST1.16 {d0, d1}, [r0]
 #define vst1q_s16 STORE_SI128
 
-void vst1q_s32(__transfersize(4) int32_t * ptr, int32x4_t val); // VST1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_s32(__transfersize(4) int32_t * ptr, int32x4_t val); // VST1.32 {d0, d1}, [r0]
 #define vst1q_s32 STORE_SI128
 
-void vst1q_s64(__transfersize(2) int64_t * ptr, int64x2_t val); // VST1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_s64(__transfersize(2) int64_t * ptr, int64x2_t val); // VST1.64 {d0, d1}, [r0]
 #define vst1q_s64 STORE_SI128
 
-void vst1q_f16(__transfersize(8) __fp16 * ptr, float16x8_t val); // VST1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_f16(__transfersize(8) __fp16 * ptr, float16x8_t val); // VST1.16 {d0, d1}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently
 
-void vst1q_f32(__transfersize(4) float32_t * ptr, float32x4_t val); // VST1.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_f32(__transfersize(4) float32_t * ptr, float32x4_t val); // VST1.32 {d0, d1}, [r0]
 _NEON2SSE_INLINE void vst1q_f32(__transfersize(4) float32_t * ptr, float32x4_t val)
 {
     if( ((uintptr_t)(ptr) & 15)  == 0 ) //16 bits aligned
@@ -9602,13 +9604,13 @@ _NEON2SSE_INLINE void vst1q_f32(__transfersize(4) float32_t * ptr, float32x4_t v
         _mm_storeu_ps (ptr, val);
 }
 
-void vst1q_p8(__transfersize(16) poly8_t * ptr, poly8x16_t val); // VST1.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_p8(__transfersize(16) poly8_t * ptr, poly8x16_t val); // VST1.8 {d0, d1}, [r0]
 #define vst1q_p8  vst1q_u8
 
-void vst1q_p16(__transfersize(8) poly16_t * ptr, poly16x8_t val); // VST1.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst1q_p16(__transfersize(8) poly16_t * ptr, poly16x8_t val); // VST1.16 {d0, d1}, [r0]
 #define vst1q_p16 vst1q_u16
 
-void vst1_u8(__transfersize(8) uint8_t * ptr, uint8x8_t val); // VST1.8 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_u8(__transfersize(8) uint8_t * ptr, uint8x8_t val); // VST1.8 {d0}, [r0]
 _NEON2SSE_INLINE void vst1_u8(__transfersize(8) uint8_t * ptr, uint8x8_t val)
 {
     int i;
@@ -9619,7 +9621,7 @@ _NEON2SSE_INLINE void vst1_u8(__transfersize(8) uint8_t * ptr, uint8x8_t val)
     return;
 }
 
-void vst1_u16(__transfersize(4) uint16_t * ptr, uint16x4_t val); // VST1.16 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_u16(__transfersize(4) uint16_t * ptr, uint16x4_t val); // VST1.16 {d0}, [r0]
 _NEON2SSE_INLINE void vst1_u16(__transfersize(4) uint16_t * ptr, uint16x4_t val)
 {
     int i;
@@ -9630,7 +9632,7 @@ _NEON2SSE_INLINE void vst1_u16(__transfersize(4) uint16_t * ptr, uint16x4_t val)
     return;
 }
 
-void vst1_u32(__transfersize(2) uint32_t * ptr, uint32x2_t val); // VST1.32 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_u32(__transfersize(2) uint32_t * ptr, uint32x2_t val); // VST1.32 {d0}, [r0]
 _NEON2SSE_INLINE void vst1_u32(__transfersize(2) uint32_t * ptr, uint32x2_t val)
 {
     int i;
@@ -9641,7 +9643,7 @@ _NEON2SSE_INLINE void vst1_u32(__transfersize(2) uint32_t * ptr, uint32x2_t val)
     return;
 }
 
-void vst1_u64(__transfersize(1) uint64_t * ptr, uint64x1_t val); // VST1.64 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_u64(__transfersize(1) uint64_t * ptr, uint64x1_t val); // VST1.64 {d0}, [r0]
 _NEON2SSE_INLINE void vst1_u64(__transfersize(1) uint64_t * ptr, uint64x1_t val)
 {
     *(ptr) = *((uint64_t*)&val);
@@ -9649,22 +9651,22 @@ _NEON2SSE_INLINE void vst1_u64(__transfersize(1) uint64_t * ptr, uint64x1_t val)
     return;
 }
 
-void vst1_s8(__transfersize(8) int8_t * ptr, int8x8_t val); // VST1.8 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_s8(__transfersize(8) int8_t * ptr, int8x8_t val); // VST1.8 {d0}, [r0]
 #define vst1_s8(ptr,val) vst1_u8((uint8_t*)ptr,val)
 
-void vst1_s16(__transfersize(4) int16_t * ptr, int16x4_t val); // VST1.16 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_s16(__transfersize(4) int16_t * ptr, int16x4_t val); // VST1.16 {d0}, [r0]
 #define vst1_s16(ptr,val) vst1_u16((uint16_t*)ptr,val)
 
-void vst1_s32(__transfersize(2) int32_t * ptr, int32x2_t val); // VST1.32 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_s32(__transfersize(2) int32_t * ptr, int32x2_t val); // VST1.32 {d0}, [r0]
 #define vst1_s32(ptr,val) vst1_u32((uint32_t*)ptr,val)
 
-void vst1_s64(__transfersize(1) int64_t * ptr, int64x1_t val); // VST1.64 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_s64(__transfersize(1) int64_t * ptr, int64x1_t val); // VST1.64 {d0}, [r0]
 #define vst1_s64(ptr,val) vst1_u64((uint64_t*)ptr,val)
 
-void vst1_f16(__transfersize(4) __fp16 * ptr, float16x4_t val); // VST1.16 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_f16(__transfersize(4) __fp16 * ptr, float16x4_t val); // VST1.16 {d0}, [r0]
 //current IA SIMD doesn't support float16
 
-void vst1_f32(__transfersize(2) float32_t * ptr, float32x2_t val); // VST1.32 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_f32(__transfersize(2) float32_t * ptr, float32x2_t val); // VST1.32 {d0}, [r0]
 _NEON2SSE_INLINE void vst1_f32(__transfersize(2) float32_t * ptr, float32x2_t val)
 {
     *(ptr) =   val.m64_f32[0];
@@ -9672,42 +9674,42 @@ _NEON2SSE_INLINE void vst1_f32(__transfersize(2) float32_t * ptr, float32x2_t va
     return;
 }
 
-void vst1_p8(__transfersize(8) poly8_t * ptr, poly8x8_t val); // VST1.8 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_p8(__transfersize(8) poly8_t * ptr, poly8x8_t val); // VST1.8 {d0}, [r0]
 #define vst1_p8 vst1_u8
 
-void vst1_p16(__transfersize(4) poly16_t * ptr, poly16x4_t val); // VST1.16 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_p16(__transfersize(4) poly16_t * ptr, poly16x4_t val); // VST1.16 {d0}, [r0]
 #define vst1_p16 vst1_u16
 
 //***********Store a lane of a vector into memory (extract given lane) *********************
 //******************************************************************************************
-void vst1q_lane_u8(__transfersize(1) uint8_t * ptr, uint8x16_t val, __constrange(0,15) int lane); // VST1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_u8(__transfersize(1) uint8_t * ptr, uint8x16_t val, __constrange(0,15) int lane); // VST1.8 {d0[0]}, [r0]
 #define vst1q_lane_u8(ptr, val, lane) *(ptr) = (uint8_t) _MM_EXTRACT_EPI8 (val, lane)
 
-void vst1q_lane_u16(__transfersize(1) uint16_t * ptr, uint16x8_t val, __constrange(0,7) int lane); // VST1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_u16(__transfersize(1) uint16_t * ptr, uint16x8_t val, __constrange(0,7) int lane); // VST1.16 {d0[0]}, [r0]
 #define vst1q_lane_u16(ptr, val, lane) *(ptr) = (uint16_t) _MM_EXTRACT_EPI16 (val, lane)
 
-void vst1q_lane_u32(__transfersize(1) uint32_t * ptr, uint32x4_t val, __constrange(0,3) int lane); // VST1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_u32(__transfersize(1) uint32_t * ptr, uint32x4_t val, __constrange(0,3) int lane); // VST1.32 {d0[0]}, [r0]
 #define vst1q_lane_u32(ptr, val, lane) *(ptr) = (uint32_t) _MM_EXTRACT_EPI32 (val, lane)
 
-void vst1q_lane_u64(__transfersize(1) uint64_t * ptr, uint64x2_t val, __constrange(0,1) int lane); // VST1.64 {d0}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_u64(__transfersize(1) uint64_t * ptr, uint64x2_t val, __constrange(0,1) int lane); // VST1.64 {d0}, [r0]
 #define vst1q_lane_u64(ptr, val, lane) *(ptr) = (uint64_t) _MM_EXTRACT_EPI64 (val, lane)
 
-void vst1q_lane_s8(__transfersize(1) int8_t * ptr, int8x16_t val, __constrange(0,15) int lane); // VST1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_s8(__transfersize(1) int8_t * ptr, int8x16_t val, __constrange(0,15) int lane); // VST1.8 {d0[0]}, [r0]
 #define vst1q_lane_s8(ptr, val, lane) *(ptr) = (int8_t) _MM_EXTRACT_EPI8 (val, lane)
 
-void vst1q_lane_s16(__transfersize(1) int16_t * ptr, int16x8_t val, __constrange(0,7) int lane); // VST1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_s16(__transfersize(1) int16_t * ptr, int16x8_t val, __constrange(0,7) int lane); // VST1.16 {d0[0]}, [r0]
 #define vst1q_lane_s16(ptr, val, lane) *(ptr) = (int16_t) _MM_EXTRACT_EPI16 (val, lane)
 
-void vst1q_lane_s32(__transfersize(1) int32_t * ptr, int32x4_t val, __constrange(0,3) int lane); // VST1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_s32(__transfersize(1) int32_t * ptr, int32x4_t val, __constrange(0,3) int lane); // VST1.32 {d0[0]}, [r0]
 #define vst1q_lane_s32(ptr, val, lane) *(ptr) = _MM_EXTRACT_EPI32 (val, lane)
 
-void vst1q_lane_s64(__transfersize(1) int64_t * ptr, int64x2_t val, __constrange(0,1) int lane); // VST1.64 {d0}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_s64(__transfersize(1) int64_t * ptr, int64x2_t val, __constrange(0,1) int lane); // VST1.64 {d0}, [r0]
 #define vst1q_lane_s64(ptr, val, lane) *(ptr) = _MM_EXTRACT_EPI64 (val, lane)
 
-void vst1q_lane_f16(__transfersize(1) __fp16 * ptr, float16x8_t val, __constrange(0,7) int lane); // VST1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_f16(__transfersize(1) __fp16 * ptr, float16x8_t val, __constrange(0,7) int lane); // VST1.16 {d0[0]}, [r0]
 //current IA SIMD doesn't support float16
 
-void vst1q_lane_f32(__transfersize(1) float32_t * ptr, float32x4_t val, __constrange(0,3) int lane); // VST1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_f32(__transfersize(1) float32_t * ptr, float32x4_t val, __constrange(0,3) int lane); // VST1.32 {d0[0]}, [r0]
 _NEON2SSE_INLINE void vst1q_lane_f32(__transfersize(1) float32_t * ptr, float32x4_t val, __constrange(0,3) int lane)
 {
     int32_t ilane;
@@ -9715,63 +9717,63 @@ _NEON2SSE_INLINE void vst1q_lane_f32(__transfersize(1) float32_t * ptr, float32x
     *(ptr) =  *((float*)&ilane);
 }
 
-void vst1q_lane_p8(__transfersize(1) poly8_t * ptr, poly8x16_t val, __constrange(0,15) int lane); // VST1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_p8(__transfersize(1) poly8_t * ptr, poly8x16_t val, __constrange(0,15) int lane); // VST1.8 {d0[0]}, [r0]
 #define vst1q_lane_p8   vst1q_lane_u8
 
-void vst1q_lane_p16(__transfersize(1) poly16_t * ptr, poly16x8_t val, __constrange(0,7) int lane); // VST1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1q_lane_p16(__transfersize(1) poly16_t * ptr, poly16x8_t val, __constrange(0,7) int lane); // VST1.16 {d0[0]}, [r0]
 #define vst1q_lane_p16   vst1q_lane_s16
 
-void vst1_lane_u8(__transfersize(1) uint8_t * ptr, uint8x8_t val, __constrange(0,7) int lane); // VST1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1_lane_u8(__transfersize(1) uint8_t * ptr, uint8x8_t val, __constrange(0,7) int lane); // VST1.8 {d0[0]}, [r0]
 _NEON2SSE_INLINE void vst1_lane_u8(__transfersize(1) uint8_t * ptr, uint8x8_t val, __constrange(0,7) int lane)
 {
     *(ptr) = val.m64_u8[lane];
 }
 
-void vst1_lane_u16(__transfersize(1) uint16_t * ptr, uint16x4_t val, __constrange(0,3) int lane); // VST1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1_lane_u16(__transfersize(1) uint16_t * ptr, uint16x4_t val, __constrange(0,3) int lane); // VST1.16 {d0[0]}, [r0]
 _NEON2SSE_INLINE void vst1_lane_u16(__transfersize(1) uint16_t * ptr, uint16x4_t val, __constrange(0,3) int lane)
 {
     *(ptr) = val.m64_u16[lane];
 }
 
-void vst1_lane_u32(__transfersize(1) uint32_t * ptr, uint32x2_t val, __constrange(0,1) int lane); // VST1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1_lane_u32(__transfersize(1) uint32_t * ptr, uint32x2_t val, __constrange(0,1) int lane); // VST1.32 {d0[0]}, [r0]
 _NEON2SSE_INLINE void vst1_lane_u32(__transfersize(1) uint32_t * ptr, uint32x2_t val, __constrange(0,1) int lane)
 {
     *(ptr) = val.m64_u32[lane];
 }
 
-void vst1_lane_u64(__transfersize(1) uint64_t * ptr, uint64x1_t val, __constrange(0,0) int lane); // VST1.64 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_lane_u64(__transfersize(1) uint64_t * ptr, uint64x1_t val, __constrange(0,0) int lane); // VST1.64 {d0}, [r0]
 _NEON2SSE_INLINE void vst1_lane_u64(__transfersize(1) uint64_t * ptr, uint64x1_t val, __constrange(0,0) int lane)
 {
     *(ptr) = val.m64_u64[0];
 }
 
-void vst1_lane_s8(__transfersize(1) int8_t * ptr, int8x8_t val, __constrange(0,7) int lane); // VST1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1_lane_s8(__transfersize(1) int8_t * ptr, int8x8_t val, __constrange(0,7) int lane); // VST1.8 {d0[0]}, [r0]
 #define  vst1_lane_s8(ptr, val, lane) vst1_lane_u8((uint8_t*)ptr, val, lane)
 
-void vst1_lane_s16(__transfersize(1) int16_t * ptr, int16x4_t val, __constrange(0,3) int lane); // VST1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1_lane_s16(__transfersize(1) int16_t * ptr, int16x4_t val, __constrange(0,3) int lane); // VST1.16 {d0[0]}, [r0]
 #define vst1_lane_s16(ptr, val, lane) vst1_lane_u16((uint16_t*)ptr, val, lane)
 
-void vst1_lane_s32(__transfersize(1) int32_t * ptr, int32x2_t val, __constrange(0,1) int lane); // VST1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1_lane_s32(__transfersize(1) int32_t * ptr, int32x2_t val, __constrange(0,1) int lane); // VST1.32 {d0[0]}, [r0]
 #define vst1_lane_s32(ptr, val, lane)  vst1_lane_u32((uint32_t*)ptr, val, lane)
 
 
-void vst1_lane_s64(__transfersize(1) int64_t * ptr, int64x1_t val, __constrange(0,0) int lane); // VST1.64 {d0}, [r0]
+_NEON2SSESTORAGE void vst1_lane_s64(__transfersize(1) int64_t * ptr, int64x1_t val, __constrange(0,0) int lane); // VST1.64 {d0}, [r0]
 #define vst1_lane_s64(ptr, val, lane) vst1_lane_u64((uint64_t*)ptr, val, lane)
 
 
-void vst1_lane_f16(__transfersize(1) __fp16 * ptr, float16x4_t val, __constrange(0,3) int lane); // VST1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1_lane_f16(__transfersize(1) __fp16 * ptr, float16x4_t val, __constrange(0,3) int lane); // VST1.16 {d0[0]}, [r0]
 //current IA SIMD doesn't support float16
 
-void vst1_lane_f32(__transfersize(1) float32_t * ptr, float32x2_t val, __constrange(0,1) int lane); // VST1.32 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1_lane_f32(__transfersize(1) float32_t * ptr, float32x2_t val, __constrange(0,1) int lane); // VST1.32 {d0[0]}, [r0]
 _NEON2SSE_INLINE void vst1_lane_f32(__transfersize(1) float32_t * ptr, float32x2_t val, __constrange(0,1) int lane)
 {
     *(ptr) = val.m64_f32[lane];
 }
 
-void vst1_lane_p8(__transfersize(1) poly8_t * ptr, poly8x8_t val, __constrange(0,7) int lane); // VST1.8 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1_lane_p8(__transfersize(1) poly8_t * ptr, poly8x8_t val, __constrange(0,7) int lane); // VST1.8 {d0[0]}, [r0]
 #define vst1_lane_p8 vst1_lane_u8
 
-void vst1_lane_p16(__transfersize(1) poly16_t * ptr, poly16x4_t val, __constrange(0,3) int lane); // VST1.16 {d0[0]}, [r0]
+_NEON2SSESTORAGE void vst1_lane_p16(__transfersize(1) poly16_t * ptr, poly16x4_t val, __constrange(0,3) int lane); // VST1.16 {d0[0]}, [r0]
 #define vst1_lane_p16 vst1_lane_s16
 
 //***********************************************************************************************
@@ -9780,7 +9782,7 @@ void vst1_lane_p16(__transfersize(1) poly16_t * ptr, poly16x4_t val, __constrang
 //These intrinsics load or store an n-element structure. The array structures are defined in the beginning
 //We assume ptr is NOT aligned in general case, for more details see  "Loads and stores of a single vector functions"
 //****************** 2 elements load  *********************************************
-uint8x16x2_t vld2q_u8(__transfersize(32) uint8_t const * ptr); // VLD2.8 {d0, d2}, [r0]
+_NEON2SSESTORAGE uint8x16x2_t vld2q_u8(__transfersize(32) uint8_t const * ptr); // VLD2.8 {d0, d2}, [r0]
 _NEON2SSE_INLINE uint8x16x2_t vld2q_u8(__transfersize(32) uint8_t const * ptr) // VLD2.8 {d0, d2}, [r0]
 {
     uint8x16x2_t v;
@@ -9790,7 +9792,7 @@ _NEON2SSE_INLINE uint8x16x2_t vld2q_u8(__transfersize(32) uint8_t const * ptr) /
     return v;
 }
 
-uint16x8x2_t vld2q_u16(__transfersize(16) uint16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE uint16x8x2_t vld2q_u16(__transfersize(16) uint16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
 _NEON2SSE_INLINE uint16x8x2_t vld2q_u16(__transfersize(16) uint16_t const * ptr) // VLD2.16 {d0, d2}, [r0]
 {
     uint16x8x2_t v;
@@ -9800,7 +9802,7 @@ _NEON2SSE_INLINE uint16x8x2_t vld2q_u16(__transfersize(16) uint16_t const * ptr)
     return v;
 }
 
-uint32x4x2_t vld2q_u32(__transfersize(8) uint32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
+_NEON2SSESTORAGE uint32x4x2_t vld2q_u32(__transfersize(8) uint32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
 _NEON2SSE_INLINE uint32x4x2_t vld2q_u32(__transfersize(8) uint32_t const * ptr) // VLD2.32 {d0, d2}, [r0]
 {
     uint32x4x2_t v;
@@ -9810,20 +9812,20 @@ _NEON2SSE_INLINE uint32x4x2_t vld2q_u32(__transfersize(8) uint32_t const * ptr)
     return v;
 }
 
-int8x16x2_t vld2q_s8(__transfersize(32) int8_t const * ptr);
+_NEON2SSESTORAGE int8x16x2_t vld2q_s8(__transfersize(32) int8_t const * ptr);
 #define  vld2q_s8(ptr) vld2q_u8((uint8_t*) ptr)
 
-int16x8x2_t vld2q_s16(__transfersize(16) int16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE int16x8x2_t vld2q_s16(__transfersize(16) int16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
 #define vld2q_s16(ptr) vld2q_u16((uint16_t*) ptr)
 
-int32x4x2_t vld2q_s32(__transfersize(8) int32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
+_NEON2SSESTORAGE int32x4x2_t vld2q_s32(__transfersize(8) int32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
 #define vld2q_s32(ptr) vld2q_u32((uint32_t*) ptr)
 
 
-float16x8x2_t vld2q_f16(__transfersize(16) __fp16 const * ptr); // VLD2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE float16x8x2_t vld2q_f16(__transfersize(16) __fp16 const * ptr); // VLD2.16 {d0, d2}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers. See vld1q_f16 for example
 
-float32x4x2_t vld2q_f32(__transfersize(8) float32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
+_NEON2SSESTORAGE float32x4x2_t vld2q_f32(__transfersize(8) float32_t const * ptr); // VLD2.32 {d0, d2}, [r0]
 _NEON2SSE_INLINE float32x4x2_t vld2q_f32(__transfersize(8) float32_t const * ptr) // VLD2.32 {d0, d2}, [r0]
 {
     float32x4x2_t v;
@@ -9833,13 +9835,13 @@ _NEON2SSE_INLINE float32x4x2_t vld2q_f32(__transfersize(8) float32_t const * ptr
     return v;
 }
 
-poly8x16x2_t vld2q_p8(__transfersize(32) poly8_t const * ptr); // VLD2.8 {d0, d2}, [r0]
+_NEON2SSESTORAGE poly8x16x2_t vld2q_p8(__transfersize(32) poly8_t const * ptr); // VLD2.8 {d0, d2}, [r0]
 #define  vld2q_p8 vld2q_u8
 
-poly16x8x2_t vld2q_p16(__transfersize(16) poly16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
+_NEON2SSESTORAGE poly16x8x2_t vld2q_p16(__transfersize(16) poly16_t const * ptr); // VLD2.16 {d0, d2}, [r0]
 #define vld2q_p16 vld2q_u16
 
-uint8x8x2_t vld2_u8(__transfersize(16) uint8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint8x8x2_t vld2_u8(__transfersize(16) uint8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
 _NEON2SSE_INLINE uint8x8x2_t vld2_u8(__transfersize(16) uint8_t const * ptr)
 {
     uint8x8x2_t v;
@@ -9850,7 +9852,7 @@ _NEON2SSE_INLINE uint8x8x2_t vld2_u8(__transfersize(16) uint8_t const * ptr)
     return v;
 }
 
-uint16x4x2_t vld2_u16(__transfersize(8) uint16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint16x4x2_t vld2_u16(__transfersize(8) uint16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
 _NEON2SSE_INLINE uint16x4x2_t vld2_u16(__transfersize(8) uint16_t const * ptr)
 {
     _NEON2SSE_ALIGN_16 uint16x4x2_t v;
@@ -9861,7 +9863,7 @@ _NEON2SSE_INLINE uint16x4x2_t vld2_u16(__transfersize(8) uint16_t const * ptr)
     return v;
 }
 
-uint32x2x2_t vld2_u32(__transfersize(4) uint32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint32x2x2_t vld2_u32(__transfersize(4) uint32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
 _NEON2SSE_INLINE uint32x2x2_t vld2_u32(__transfersize(4) uint32_t const * ptr)
 {
     _NEON2SSE_ALIGN_16 uint32x2x2_t v;
@@ -9872,7 +9874,7 @@ _NEON2SSE_INLINE uint32x2x2_t vld2_u32(__transfersize(4) uint32_t const * ptr)
     return v;
 }
 
-uint64x1x2_t vld2_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint64x1x2_t vld2_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
 _NEON2SSE_INLINE uint64x1x2_t vld2_u64(__transfersize(2) uint64_t const * ptr)
 {
     uint64x1x2_t v;
@@ -9881,22 +9883,22 @@ _NEON2SSE_INLINE uint64x1x2_t vld2_u64(__transfersize(2) uint64_t const * ptr)
     return v;
 }
 
-int8x8x2_t vld2_s8(__transfersize(16) int8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE int8x8x2_t vld2_s8(__transfersize(16) int8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
 #define vld2_s8(ptr) vld2_u8((uint8_t*)ptr)
 
-int16x4x2_t vld2_s16(__transfersize(8) int16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE int16x4x2_t vld2_s16(__transfersize(8) int16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
 #define vld2_s16(ptr) vld2_u16((uint16_t*)ptr)
 
-int32x2x2_t vld2_s32(__transfersize(4) int32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE int32x2x2_t vld2_s32(__transfersize(4) int32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
 #define vld2_s32(ptr) vld2_u32((uint32_t*)ptr)
 
-int64x1x2_t vld2_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE int64x1x2_t vld2_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
 #define vld2_s64(ptr) vld2_u64((uint64_t*)ptr)
 
-float16x4x2_t vld2_f16(__transfersize(8) __fp16 const * ptr); // VLD2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE float16x4x2_t vld2_f16(__transfersize(8) __fp16 const * ptr); // VLD2.16 {d0, d1}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers. See vld1_f16 for example
 
-float32x2x2_t vld2_f32(__transfersize(4) float32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE float32x2x2_t vld2_f32(__transfersize(4) float32_t const * ptr); // VLD2.32 {d0, d1}, [r0]
 _NEON2SSE_INLINE float32x2x2_t vld2_f32(__transfersize(4) float32_t const * ptr)
 {
     float32x2x2_t v;
@@ -9907,15 +9909,15 @@ _NEON2SSE_INLINE float32x2x2_t vld2_f32(__transfersize(4) float32_t const * ptr)
     return v;
 }
 
-poly8x8x2_t vld2_p8(__transfersize(16) poly8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE poly8x8x2_t vld2_p8(__transfersize(16) poly8_t const * ptr); // VLD2.8 {d0, d1}, [r0]
 #define vld2_p8 vld2_u8
 
-poly16x4x2_t vld2_p16(__transfersize(8) poly16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE poly16x4x2_t vld2_p16(__transfersize(8) poly16_t const * ptr); // VLD2.16 {d0, d1}, [r0]
 #define vld2_p16 vld2_u16
 
 //******************** Triplets ***************************************
 //*********************************************************************
-uint8x16x3_t vld3q_u8(__transfersize(48) uint8_t const * ptr); // VLD3.8 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE uint8x16x3_t vld3q_u8(__transfersize(48) uint8_t const * ptr); // VLD3.8 {d0, d2, d4}, [r0]
 _NEON2SSE_INLINE uint8x16x3_t vld3q_u8(__transfersize(48) uint8_t const * ptr) // VLD3.8 {d0, d2, d4}, [r0]
 {
     //a0,a1,a2,a3,...a7,a8,...a15,  b0,b1,b2,...b7,b8,...b15, c0,c1,c2,...c7,c8,...c15 ->
@@ -9924,9 +9926,9 @@ _NEON2SSE_INLINE uint8x16x3_t vld3q_u8(__transfersize(48) uint8_t const * ptr) /
     //a:2,5,8,11,14,  b:1,4,7,10,13,  c:0,3,6,9,12,15
     uint8x16x3_t v;
     __m128i tmp0, tmp1,tmp2, tmp3;
-    _NEON2SSE_ALIGN_16 int8_t mask8_0[16] = {0,3,6,9,12,15,1,4,7,10,13,2,5,8,11,14};
-    _NEON2SSE_ALIGN_16 int8_t mask8_1[16] = {2,5,8,11,14,0,3,6,9,12,15,1,4,7,10,13};
-    _NEON2SSE_ALIGN_16 int8_t mask8_2[16] = {1,4,7,10,13,2,5,8,11,14,0,3,6,9,12,15};
+    _NEON2SSE_ALIGN_16 static const int8_t mask8_0[16] = {0,3,6,9,12,15,1,4,7,10,13,2,5,8,11,14};
+    _NEON2SSE_ALIGN_16 static const int8_t mask8_1[16] = {2,5,8,11,14,0,3,6,9,12,15,1,4,7,10,13};
+    _NEON2SSE_ALIGN_16 static const int8_t mask8_2[16] = {1,4,7,10,13,2,5,8,11,14,0,3,6,9,12,15};
 
     v.val[0] =  vld1q_u8 (ptr); //a0,a1,a2,a3,...a7, ...a15
     v.val[1] =  vld1q_u8 ((ptr + 16)); //b0,b1,b2,b3...b7, ...b15
@@ -9964,15 +9966,15 @@ _NEON2SSE_INLINE uint8x16x3_t vld3q_u8(__transfersize(48) uint8_t const * ptr) /
     return v;
 }
 
-uint16x8x3_t vld3q_u16(__transfersize(24) uint16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE uint16x8x3_t vld3q_u16(__transfersize(24) uint16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
 _NEON2SSE_INLINE uint16x8x3_t vld3q_u16(__transfersize(24) uint16_t const * ptr) // VLD3.16 {d0, d2, d4}, [r0]
 {
     //a0, a1,a2,a3,...a7,  b0,b1,b2,b3,...b7, c0,c1,c2,c3...c7 -> a0,a3,a6,b1,b4,b7,c2,c5, a1,a4,a7,b2,b5,c0,c3,c6, a2,a5,b0,b3,b6,c1,c4,c7
     uint16x8x3_t v;
     __m128i tmp0, tmp1,tmp2, tmp3;
-    _NEON2SSE_ALIGN_16 int8_t mask16_0[16] = {0,1, 6,7, 12,13, 2,3, 8,9, 14,15, 4,5, 10,11};
-    _NEON2SSE_ALIGN_16 int8_t mask16_1[16] = {2,3, 8,9, 14,15, 4,5, 10,11, 0,1, 6,7, 12,13};
-    _NEON2SSE_ALIGN_16 int8_t mask16_2[16] = {4,5, 10,11, 0,1, 6,7, 12,13, 2,3, 8,9, 14,15};
+    _NEON2SSE_ALIGN_16 static const int8_t mask16_0[16] = {0,1, 6,7, 12,13, 2,3, 8,9, 14,15, 4,5, 10,11};
+    _NEON2SSE_ALIGN_16 static const int8_t mask16_1[16] = {2,3, 8,9, 14,15, 4,5, 10,11, 0,1, 6,7, 12,13};
+    _NEON2SSE_ALIGN_16 static const int8_t mask16_2[16] = {4,5, 10,11, 0,1, 6,7, 12,13, 2,3, 8,9, 14,15};
 
     v.val[0] =  vld1q_u16 (ptr); //a0,a1,a2,a3,...a7,
     v.val[1] =  vld1q_u16 ((ptr + 8)); //b0,b1,b2,b3...b7
@@ -10010,7 +10012,7 @@ _NEON2SSE_INLINE uint16x8x3_t vld3q_u16(__transfersize(24) uint16_t const * ptr)
     return v;
 }
 
-uint32x4x3_t vld3q_u32(__transfersize(12) uint32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE uint32x4x3_t vld3q_u32(__transfersize(12) uint32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
 _NEON2SSE_INLINE uint32x4x3_t vld3q_u32(__transfersize(12) uint32_t const * ptr) // VLD3.32 {d0, d2, d4}, [r0]
 {
     //a0,a1,a2,a3,  b0,b1,b2,b3, c0,c1,c2,c3 -> a0,a3,b2,c1,  a1,b0,b3,c2, a2,b1,c0,c3,
@@ -10033,19 +10035,19 @@ _NEON2SSE_INLINE uint32x4x3_t vld3q_u32(__transfersize(12) uint32_t const * ptr)
     return v;
 }
 
-int8x16x3_t vld3q_s8(__transfersize(48) int8_t const * ptr); // VLD3.8 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE int8x16x3_t vld3q_s8(__transfersize(48) int8_t const * ptr); // VLD3.8 {d0, d2, d4}, [r0]
 #define  vld3q_s8(ptr) vld3q_u8((uint8_t*) (ptr))
 
-int16x8x3_t vld3q_s16(__transfersize(24) int16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE int16x8x3_t vld3q_s16(__transfersize(24) int16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
 #define  vld3q_s16(ptr) vld3q_u16((uint16_t*) (ptr))
 
-int32x4x3_t vld3q_s32(__transfersize(12) int32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE int32x4x3_t vld3q_s32(__transfersize(12) int32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
 #define  vld3q_s32(ptr) vld3q_u32((uint32_t*) (ptr))
 
-float16x8x3_t vld3q_f16(__transfersize(24) __fp16 const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE float16x8x3_t vld3q_f16(__transfersize(24) __fp16 const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers. See vld1q_f16 for example
 
-float32x4x3_t vld3q_f32(__transfersize(12) float32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE float32x4x3_t vld3q_f32(__transfersize(12) float32_t const * ptr); // VLD3.32 {d0, d2, d4}, [r0]
 _NEON2SSE_INLINE float32x4x3_t vld3q_f32(__transfersize(12) float32_t const * ptr) // VLD3.32 {d0, d2, d4}, [r0]
 {
     //a0,a1,a2,a3,  b0,b1,b2,b3, c0,c1,c2,c3 -> a0,a3,b2,c1,  a1,b0,b3,c2, a2,b1,c0,c3,
@@ -10071,17 +10073,17 @@ _NEON2SSE_INLINE float32x4x3_t vld3q_f32(__transfersize(12) float32_t const * pt
 poly8x16x3_t vld3q_p8(__transfersize(48) poly8_t const * ptr); // VLD3.8 {d0, d2, d4}, [r0]
 #define vld3q_p8 vld3q_u8
 
-poly16x8x3_t vld3q_p16(__transfersize(24) poly16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
+_NEON2SSESTORAGE poly16x8x3_t vld3q_p16(__transfersize(24) poly16_t const * ptr); // VLD3.16 {d0, d2, d4}, [r0]
 #define vld3q_p16 vld3q_u16
 
-uint8x8x3_t vld3_u8(__transfersize(24) uint8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE uint8x8x3_t vld3_u8(__transfersize(24) uint8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
 _NEON2SSE_INLINE uint8x8x3_t vld3_u8(__transfersize(24) uint8_t const * ptr) // VLD3.8 {d0, d1, d2}, [r0]
 {
     //a0, a1,a2,a3,...a7,  b0,b1,b2,b3,...b7, c0,c1,c2,c3...c7 -> a0,a3,a6,b1,b4,b7,c2,c5, a1,a4,a7,b2,b5,c0,c3,c6, a2,a5,b0,b3,b6,c1,c4,c7
     uint8x8x3_t v;
     __m128i val0, val1, val2, tmp0, tmp1;
-    _NEON2SSE_ALIGN_16 int8_t mask8_0[16] = {0,3,6,9,12,15, 1,4,7,10,13, 2,5,8,11,14};
-    _NEON2SSE_ALIGN_16 int8_t mask8_1[16] = {2,5, 0,3,6, 1,4,7, 0,0,0,0,0,0,0,0};
+    _NEON2SSE_ALIGN_16 static const int8_t mask8_0[16] = {0,3,6,9,12,15, 1,4,7,10,13, 2,5,8,11,14};
+    _NEON2SSE_ALIGN_16 static const int8_t mask8_1[16] = {2,5, 0,3,6, 1,4,7, 0,0,0,0,0,0,0,0};
     val0 =  vld1q_u8 (ptr); //a0,a1,a2,a3,...a7, b0,b1,b2,b3...b7
     val2 =  _mm_loadl_epi64((__m128i*)(ptr + 16)); //c0,c1,c2,c3,...c7
 
@@ -10107,13 +10109,13 @@ _NEON2SSE_INLINE uint8x8x3_t vld3_u8(__transfersize(24) uint8_t const * ptr) //
     return v;
 }
 
-uint16x4x3_t vld3_u16(__transfersize(12) uint16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE uint16x4x3_t vld3_u16(__transfersize(12) uint16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
 _NEON2SSE_INLINE uint16x4x3_t vld3_u16(__transfersize(12) uint16_t const * ptr) // VLD3.16 {d0, d1, d2}, [r0]
 {
     //a0,a1,a2,a3,  b0,b1,b2,b3, c0,c1,c2,c3 -> a0,a3,b2,c1,  a1,b0,b3,c2, a2,b1,c0,c3,
     uint16x4x3_t v;
     __m128i val0, val1, val2, tmp0, tmp1;
-    _NEON2SSE_ALIGN_16 int8_t mask16[16] = {0,1, 6,7, 12,13, 2,3, 8,9, 14,15, 4,5, 10,11};
+    _NEON2SSE_ALIGN_16 static const int8_t mask16[16] = {0,1, 6,7, 12,13, 2,3, 8,9, 14,15, 4,5, 10,11};
     val0 =  vld1q_u16 (ptr); //a0,a1,a2,a3,  b0,b1,b2,b3
     val2 =  _mm_loadl_epi64((__m128i*)(ptr + 8)); //c0,c1,c2,c3, x,x,x,x
 
@@ -10141,7 +10143,7 @@ _NEON2SSE_INLINE uint16x4x3_t vld3_u16(__transfersize(12) uint16_t const * ptr)
     return v;
 }
 
-uint32x2x3_t vld3_u32(__transfersize(6) uint32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE uint32x2x3_t vld3_u32(__transfersize(6) uint32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
 _NEON2SSE_INLINE uint32x2x3_t vld3_u32(__transfersize(6) uint32_t const * ptr) // VLD3.32 {d0, d1, d2}, [r0]
 {
     //a0,a1,  b0,b1, c0,c1,  -> a0,b1, a1,c0, b0,c1
@@ -10159,7 +10161,7 @@ _NEON2SSE_INLINE uint32x2x3_t vld3_u32(__transfersize(6) uint32_t const * ptr) /
     _M64(v.val[2], val2);
     return v;
 }
-uint64x1x3_t vld3_u64(__transfersize(3) uint64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE uint64x1x3_t vld3_u64(__transfersize(3) uint64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
 _NEON2SSE_INLINE uint64x1x3_t vld3_u64(__transfersize(3) uint64_t const * ptr) // VLD1.64 {d0, d1, d2}, [r0]
 {
     uint64x1x3_t v;
@@ -10169,22 +10171,22 @@ _NEON2SSE_INLINE uint64x1x3_t vld3_u64(__transfersize(3) uint64_t const * ptr) /
     return v;
 }
 
-int8x8x3_t vld3_s8(__transfersize(24) int8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE int8x8x3_t vld3_s8(__transfersize(24) int8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
 #define vld3_s8(ptr) vld3_u8((uint8_t*)ptr)
 
-int16x4x3_t vld3_s16(__transfersize(12) int16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE int16x4x3_t vld3_s16(__transfersize(12) int16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
 #define vld3_s16(ptr) vld3_u16((uint16_t*)ptr)
 
-int32x2x3_t vld3_s32(__transfersize(6) int32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE int32x2x3_t vld3_s32(__transfersize(6) int32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
 #define vld3_s32(ptr) vld3_u32((uint32_t*)ptr)
 
 int64x1x3_t vld3_s64(__transfersize(3) int64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
 #define vld3_s64(ptr) vld3_u64((uint64_t*)ptr)
 
-float16x4x3_t vld3_f16(__transfersize(12) __fp16 const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE float16x4x3_t vld3_f16(__transfersize(12) __fp16 const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers. See vld1q_f16 for example
 
-float32x2x3_t vld3_f32(__transfersize(6) float32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE float32x2x3_t vld3_f32(__transfersize(6) float32_t const * ptr); // VLD3.32 {d0, d1, d2}, [r0]
 _NEON2SSE_INLINE float32x2x3_t vld3_f32(__transfersize(6) float32_t const * ptr)
 {
     //a0,a1,  b0,b1, c0,c1,  -> a0,b1, a1,c0, b0,c1
@@ -10200,15 +10202,15 @@ _NEON2SSE_INLINE float32x2x3_t vld3_f32(__transfersize(6) float32_t const * ptr)
     return v;
 }
 
-poly8x8x3_t vld3_p8(__transfersize(24) poly8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE poly8x8x3_t vld3_p8(__transfersize(24) poly8_t const * ptr); // VLD3.8 {d0, d1, d2}, [r0]
 #define vld3_p8 vld3_u8
 
-poly16x4x3_t vld3_p16(__transfersize(12) poly16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE poly16x4x3_t vld3_p16(__transfersize(12) poly16_t const * ptr); // VLD3.16 {d0, d1, d2}, [r0]
 #define vld3_p16 vld3_u16
 
 //***************  Quadruples load ********************************
 //*****************************************************************
-uint8x16x4_t vld4q_u8(__transfersize(64) uint8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE uint8x16x4_t vld4q_u8(__transfersize(64) uint8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
 _NEON2SSE_INLINE uint8x16x4_t vld4q_u8(__transfersize(64) uint8_t const * ptr) // VLD4.8 {d0, d2, d4, d6}, [r0]
 {
     uint8x16x4_t v;
@@ -10241,7 +10243,7 @@ _NEON2SSE_INLINE uint8x16x4_t vld4q_u8(__transfersize(64) uint8_t const * ptr) /
     return v;
 }
 
-uint16x8x4_t vld4q_u16(__transfersize(32) uint16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE uint16x8x4_t vld4q_u16(__transfersize(32) uint16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
 _NEON2SSE_INLINE uint16x8x4_t vld4q_u16(__transfersize(32) uint16_t const * ptr) // VLD4.16 {d0, d2, d4, d6}, [r0]
 {
     uint16x8x4_t v;
@@ -10265,7 +10267,7 @@ _NEON2SSE_INLINE uint16x8x4_t vld4q_u16(__transfersize(32) uint16_t const * ptr)
     return v;
 }
 
-uint32x4x4_t vld4q_u32(__transfersize(16) uint32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE uint32x4x4_t vld4q_u32(__transfersize(16) uint32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
 _NEON2SSE_INLINE uint32x4x4_t vld4q_u32(__transfersize(16) uint32_t const * ptr) // VLD4.32 {d0, d2, d4, d6}, [r0]
 {
     uint32x4x4_t v;
@@ -10285,19 +10287,19 @@ _NEON2SSE_INLINE uint32x4x4_t vld4q_u32(__transfersize(16) uint32_t const * ptr)
     return v;
 }
 
-int8x16x4_t vld4q_s8(__transfersize(64) int8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE int8x16x4_t vld4q_s8(__transfersize(64) int8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
 #define vld4q_s8(ptr) vld4q_u8((uint8_t*)ptr)
 
-int16x8x4_t vld4q_s16(__transfersize(32) int16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE int16x8x4_t vld4q_s16(__transfersize(32) int16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
 #define  vld4q_s16(ptr) vld4q_u16((uint16_t*)ptr)
 
-int32x4x4_t vld4q_s32(__transfersize(16) int32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE int32x4x4_t vld4q_s32(__transfersize(16) int32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
 #define  vld4q_s32(ptr) vld4q_u32((uint32_t*)ptr)
 
-float16x8x4_t vld4q_f16(__transfersize(32) __fp16 const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE float16x8x4_t vld4q_f16(__transfersize(32) __fp16 const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers. See vld1q_f16 for example
 
-float32x4x4_t vld4q_f32(__transfersize(16) float32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE float32x4x4_t vld4q_f32(__transfersize(16) float32_t const * ptr); // VLD4.32 {d0, d2, d4, d6}, [r0]
 _NEON2SSE_INLINE float32x4x4_t vld4q_f32(__transfersize(16) float32_t const * ptr) // VLD4.32 {d0, d2, d4, d6}, [r0]
 {
     float32x4x4_t v;
@@ -10318,13 +10320,13 @@ _NEON2SSE_INLINE float32x4x4_t vld4q_f32(__transfersize(16) float32_t const * pt
     return v;
 }
 
-poly8x16x4_t vld4q_p8(__transfersize(64) poly8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE poly8x16x4_t vld4q_p8(__transfersize(64) poly8_t const * ptr); // VLD4.8 {d0, d2, d4, d6}, [r0]
 #define vld4q_p8 vld4q_u8
 
-poly16x8x4_t vld4q_p16(__transfersize(32) poly16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
+_NEON2SSESTORAGE poly16x8x4_t vld4q_p16(__transfersize(32) poly16_t const * ptr); // VLD4.16 {d0, d2, d4, d6}, [r0]
 #define vld4q_p16 vld4q_s16
 
-uint8x8x4_t vld4_u8(__transfersize(32) uint8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE uint8x8x4_t vld4_u8(__transfersize(32) uint8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
 _NEON2SSE_INLINE uint8x8x4_t vld4_u8(__transfersize(32) uint8_t const * ptr) // VLD4.8 {d0, d1, d2, d3}, [r0]
 {
     uint8x8x4_t v;
@@ -10344,13 +10346,13 @@ _NEON2SSE_INLINE uint8x8x4_t vld4_u8(__transfersize(32) uint8_t const * ptr) //
     return v;
 }
 
-uint16x4x4_t vld4_u16(__transfersize(16) uint16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE uint16x4x4_t vld4_u16(__transfersize(16) uint16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
 _NEON2SSE_INLINE uint16x4x4_t vld4_u16(__transfersize(16) uint16_t const * ptr) // VLD4.16 {d0, d1, d2, d3}, [r0]
 {
     uint16x4x4_t v;
     __m128i sh0, sh1;
     __m128i val0, val2;
-    _NEON2SSE_ALIGN_16 int8_t mask4_16[16] = {0,1, 8,9, 2,3, 10,11, 4,5, 12,13, 6,7, 14,15}; //0, 4, 1, 5, 2, 6, 3, 7
+    _NEON2SSE_ALIGN_16 static const int8_t mask4_16[16] = {0,1, 8,9, 2,3, 10,11, 4,5, 12,13, 6,7, 14,15}; //0, 4, 1, 5, 2, 6, 3, 7
     val0 = vld1q_u16 ( (ptr)); //load first 64-bits in val[0] and val[1]
     val2 = vld1q_u16 ( (ptr + 8)); //load third and forth 64-bits in val[2], val[3]
     sh0 = _mm_shuffle_epi8(val0, *(__m128i*)mask4_16);
@@ -10362,7 +10364,7 @@ _NEON2SSE_INLINE uint16x4x4_t vld4_u16(__transfersize(16) uint16_t const * ptr)
     return v;
 }
 
-uint32x2x4_t vld4_u32(__transfersize(8) uint32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE uint32x2x4_t vld4_u32(__transfersize(8) uint32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
 _NEON2SSE_INLINE uint32x2x4_t vld4_u32(__transfersize(8) uint32_t const * ptr)
 {
     //a0,a1,  b0,b1, c0,c1, d0,d1 -> a0,c0, a1,c1, b0,d0, b1,d1
@@ -10377,7 +10379,7 @@ _NEON2SSE_INLINE uint32x2x4_t vld4_u32(__transfersize(8) uint32_t const * ptr)
     return v;
 }
 
-uint64x1x4_t vld4_u64(__transfersize(4) uint64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE uint64x1x4_t vld4_u64(__transfersize(4) uint64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
 _NEON2SSE_INLINE uint64x1x4_t vld4_u64(__transfersize(4) uint64_t const * ptr) // VLD1.64 {d0, d1, d2, d3}, [r0]
 {
     uint64x1x4_t v;
@@ -10388,22 +10390,22 @@ _NEON2SSE_INLINE uint64x1x4_t vld4_u64(__transfersize(4) uint64_t const * ptr) /
     return v;
 }
 
-int8x8x4_t vld4_s8(__transfersize(32) int8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE int8x8x4_t vld4_s8(__transfersize(32) int8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
 #define  vld4_s8(ptr) vld4_u8((uint8_t*)ptr)
 
-int16x4x4_t vld4_s16(__transfersize(16) int16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE int16x4x4_t vld4_s16(__transfersize(16) int16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
 #define vld4_s16(ptr) vld4_u16((uint16_t*)ptr)
 
-int32x2x4_t vld4_s32(__transfersize(8) int32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE int32x2x4_t vld4_s32(__transfersize(8) int32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
 #define vld4_s32(ptr) vld4_u32((uint32_t*)ptr)
 
 int64x1x4_t vld4_s64(__transfersize(4) int64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
 #define vld4_s64(ptr) vld4_u64((uint64_t*)ptr)
 
-float16x4x4_t vld4_f16(__transfersize(16) __fp16 const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE float16x4x4_t vld4_f16(__transfersize(16) __fp16 const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers. See vld1q_f16 for example
 
-float32x2x4_t vld4_f32(__transfersize(8) float32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE float32x2x4_t vld4_f32(__transfersize(8) float32_t const * ptr); // VLD4.32 {d0, d1, d2, d3}, [r0]
 _NEON2SSE_INLINE float32x2x4_t vld4_f32(__transfersize(8) float32_t const * ptr) // VLD4.32 {d0, d1, d2, d3}, [r0]
 {
     //a0,a1,  b0,b1, c0,c1, d0,d1 -> a0,c0, a1,c1, b0,d0, b1,d1
@@ -10419,15 +10421,15 @@ _NEON2SSE_INLINE float32x2x4_t vld4_f32(__transfersize(8) float32_t const * ptr)
     return res;
 }
 
-poly8x8x4_t vld4_p8(__transfersize(32) poly8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE poly8x8x4_t vld4_p8(__transfersize(32) poly8_t const * ptr); // VLD4.8 {d0, d1, d2, d3}, [r0]
 #define vld4_p8 vld4_u8
 
-poly16x4x4_t vld4_p16(__transfersize(16) poly16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE poly16x4x4_t vld4_p16(__transfersize(16) poly16_t const * ptr); // VLD4.16 {d0, d1, d2, d3}, [r0]
 #define vld4_p16 vld4_u16
 
 //************* Duplicate (or propagate) ptr[0] to all val[0] lanes and ptr[1] to all val[1] lanes *******************
 //*******************************************************************************************************************
-uint8x8x2_t vld2_dup_u8(__transfersize(2) uint8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE uint8x8x2_t vld2_dup_u8(__transfersize(2) uint8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
 _NEON2SSE_INLINE uint8x8x2_t vld2_dup_u8(__transfersize(2) uint8_t const * ptr) // VLD2.8 {d0[], d1[]}, [r0]
 {
     uint8x8x2_t v;
@@ -10440,7 +10442,7 @@ _NEON2SSE_INLINE uint8x8x2_t vld2_dup_u8(__transfersize(2) uint8_t const * ptr)
     return v;
 }
 
-uint16x4x2_t vld2_dup_u16(__transfersize(2) uint16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE uint16x4x2_t vld2_dup_u16(__transfersize(2) uint16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
 _NEON2SSE_INLINE uint16x4x2_t vld2_dup_u16(__transfersize(2) uint16_t const * ptr) // VLD2.16 {d0[], d1[]}, [r0]
 {
     uint16x4x2_t v;
@@ -10453,7 +10455,7 @@ _NEON2SSE_INLINE uint16x4x2_t vld2_dup_u16(__transfersize(2) uint16_t const * pt
     return v;
 }
 
-uint32x2x2_t vld2_dup_u32(__transfersize(2) uint32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE uint32x2x2_t vld2_dup_u32(__transfersize(2) uint32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
 _NEON2SSE_INLINE uint32x2x2_t vld2_dup_u32(__transfersize(2) uint32_t const * ptr) // VLD2.32 {d0[], d1[]}, [r0]
 {
     uint32x2x2_t v;
@@ -10464,25 +10466,25 @@ _NEON2SSE_INLINE uint32x2x2_t vld2_dup_u32(__transfersize(2) uint32_t const * pt
     return v;
 }
 
-uint64x1x2_t vld2_dup_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE uint64x1x2_t vld2_dup_u64(__transfersize(2) uint64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
 #define vld2_dup_u64 vld2_u64
 
-int8x8x2_t vld2_dup_s8(__transfersize(2) int8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE int8x8x2_t vld2_dup_s8(__transfersize(2) int8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
 #define vld2_dup_s8(ptr) vld2_dup_u8((uint8_t*)ptr)
 
-int16x4x2_t vld2_dup_s16(__transfersize(2) int16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE int16x4x2_t vld2_dup_s16(__transfersize(2) int16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
 #define vld2_dup_s16(ptr) vld2_dup_u16((uint16_t*)ptr)
 
-int32x2x2_t vld2_dup_s32(__transfersize(2) int32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE int32x2x2_t vld2_dup_s32(__transfersize(2) int32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
 #define vld2_dup_s32(ptr) vld2_dup_u32((uint32_t*)ptr)
 
-int64x1x2_t vld2_dup_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE int64x1x2_t vld2_dup_s64(__transfersize(2) int64_t const * ptr); // VLD1.64 {d0, d1}, [r0]
 #define vld2_dup_s64(ptr) vld2_dup_u64((uint64_t*)ptr)
 
-float16x4x2_t vld2_dup_f16(__transfersize(2) __fp16 const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE float16x4x2_t vld2_dup_f16(__transfersize(2) __fp16 const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers. See vld1q_f16 for example
 
-float32x2x2_t vld2_dup_f32(__transfersize(2) float32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE float32x2x2_t vld2_dup_f32(__transfersize(2) float32_t const * ptr); // VLD2.32 {d0[], d1[]}, [r0]
 _NEON2SSE_INLINE float32x2x2_t vld2_dup_f32(__transfersize(2) float32_t const * ptr) // VLD2.32 {d0[], d1[]}, [r0]
 {
     float32x2x2_t v;
@@ -10493,16 +10495,16 @@ _NEON2SSE_INLINE float32x2x2_t vld2_dup_f32(__transfersize(2) float32_t const *
     return v;
 }
 
-poly8x8x2_t vld2_dup_p8(__transfersize(2) poly8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE poly8x8x2_t vld2_dup_p8(__transfersize(2) poly8_t const * ptr); // VLD2.8 {d0[], d1[]}, [r0]
 #define vld2_dup_p8 vld2_dup_u8
 
-poly16x4x2_t vld2_dup_p16(__transfersize(2) poly16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
+_NEON2SSESTORAGE poly16x4x2_t vld2_dup_p16(__transfersize(2) poly16_t const * ptr); // VLD2.16 {d0[], d1[]}, [r0]
 #define vld2_dup_p16 vld2_dup_s16
 
 //************* Duplicate (or propagate)triplets: *******************
 //********************************************************************
 //ptr[0] to all val[0] lanes, ptr[1] to all val[1] lanes and ptr[2] to all val[2] lanes
-uint8x8x3_t vld3_dup_u8(__transfersize(3) uint8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE uint8x8x3_t vld3_dup_u8(__transfersize(3) uint8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
 _NEON2SSE_INLINE uint8x8x3_t vld3_dup_u8(__transfersize(3) uint8_t const * ptr) // VLD3.8 {d0[], d1[], d2[]}, [r0]
 {
     uint8x8x3_t v;
@@ -10517,7 +10519,7 @@ _NEON2SSE_INLINE uint8x8x3_t vld3_dup_u8(__transfersize(3) uint8_t const * ptr)
     return v;
 }
 
-uint16x4x3_t vld3_dup_u16(__transfersize(3) uint16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE uint16x4x3_t vld3_dup_u16(__transfersize(3) uint16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
 _NEON2SSE_INLINE uint16x4x3_t vld3_dup_u16(__transfersize(3) uint16_t const * ptr) // VLD3.16 {d0[], d1[], d2[]}, [r0]
 {
     uint16x4x3_t v;
@@ -10532,7 +10534,7 @@ _NEON2SSE_INLINE uint16x4x3_t vld3_dup_u16(__transfersize(3) uint16_t const * pt
     return v;
 }
 
-uint32x2x3_t vld3_dup_u32(__transfersize(3) uint32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE uint32x2x3_t vld3_dup_u32(__transfersize(3) uint32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
 _NEON2SSE_INLINE uint32x2x3_t vld3_dup_u32(__transfersize(3) uint32_t const * ptr) // VLD3.32 {d0[], d1[], d2[]}, [r0]
 {
     uint32x2x3_t v;
@@ -10547,7 +10549,7 @@ _NEON2SSE_INLINE uint32x2x3_t vld3_dup_u32(__transfersize(3) uint32_t const * pt
     return v;
 }
 
-uint64x1x3_t vld3_dup_u64(__transfersize(3) uint64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE uint64x1x3_t vld3_dup_u64(__transfersize(3) uint64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
 _NEON2SSE_INLINE uint64x1x3_t vld3_dup_u64(__transfersize(3) uint64_t const * ptr) // VLD1.64 {d0, d1, d2}, [r0]
 {
     uint64x1x3_t v;
@@ -10557,23 +10559,23 @@ _NEON2SSE_INLINE uint64x1x3_t vld3_dup_u64(__transfersize(3) uint64_t const * pt
     return v;
 }
 
-int8x8x3_t vld3_dup_s8(__transfersize(3) int8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE int8x8x3_t vld3_dup_s8(__transfersize(3) int8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
 #define vld3_dup_s8(ptr) vld3_dup_u8((uint8_t*)ptr)
 
-int16x4x3_t vld3_dup_s16(__transfersize(3) int16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE int16x4x3_t vld3_dup_s16(__transfersize(3) int16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
 #define vld3_dup_s16(ptr) vld3_dup_u16((uint16_t*)ptr)
 
-int32x2x3_t vld3_dup_s32(__transfersize(3) int32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE int32x2x3_t vld3_dup_s32(__transfersize(3) int32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
 #define vld3_dup_s32(ptr) vld3_dup_u32((uint32_t*)ptr)
 
 int64x1x3_t vld3_dup_s64(__transfersize(3) int64_t const * ptr); // VLD1.64 {d0, d1, d2}, [r0]
 #define vld3_dup_s64(ptr) vld3_dup_u64((uint64_t*)ptr)
 
 
-float16x4x3_t vld3_dup_f16(__transfersize(3) __fp16 const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE float16x4x3_t vld3_dup_f16(__transfersize(3) __fp16 const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers. See vld1q_f16 for example
 
-float32x2x3_t vld3_dup_f32(__transfersize(3) float32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE float32x2x3_t vld3_dup_f32(__transfersize(3) float32_t const * ptr); // VLD3.32 {d0[], d1[], d2[]}, [r0]
 _NEON2SSE_INLINE float32x2x3_t vld3_dup_f32(__transfersize(3) float32_t const * ptr) // VLD3.32 {d0[], d1[], d2[]}, [r0]
 {
     float32x2x3_t v;
@@ -10585,17 +10587,17 @@ _NEON2SSE_INLINE float32x2x3_t vld3_dup_f32(__transfersize(3) float32_t const *
     return v;
 }
 
-poly8x8x3_t vld3_dup_p8(__transfersize(3) poly8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE poly8x8x3_t vld3_dup_p8(__transfersize(3) poly8_t const * ptr); // VLD3.8 {d0[], d1[], d2[]}, [r0]
 #define vld3_dup_p8 vld3_dup_u8
 
-poly16x4x3_t vld3_dup_p16(__transfersize(3) poly16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
+_NEON2SSESTORAGE poly16x4x3_t vld3_dup_p16(__transfersize(3) poly16_t const * ptr); // VLD3.16 {d0[], d1[], d2[]}, [r0]
 #define vld3_dup_p16 vld3_dup_s16
 
 
 //************* Duplicate (or propagate) quadruples: *******************
 //***********************************************************************
 //ptr[0] to all val[0] lanes, ptr[1] to all val[1] lanes, ptr[2] to all val[2] lanes  and  ptr[3] to all val[3] lanes
-uint8x8x4_t vld4_dup_u8(__transfersize(4) uint8_t const * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE uint8x8x4_t vld4_dup_u8(__transfersize(4) uint8_t const * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
 _NEON2SSE_INLINE uint8x8x4_t vld4_dup_u8(__transfersize(4) uint8_t const * ptr) // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
 {
     uint8x8x4_t v;
@@ -10610,7 +10612,7 @@ _NEON2SSE_INLINE uint8x8x4_t vld4_dup_u8(__transfersize(4) uint8_t const * ptr)
     return v;
 }
 
-uint16x4x4_t vld4_dup_u16(__transfersize(4) uint16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE uint16x4x4_t vld4_dup_u16(__transfersize(4) uint16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
 _NEON2SSE_INLINE uint16x4x4_t vld4_dup_u16(__transfersize(4) uint16_t const * ptr) // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
 {
     uint16x4x4_t v;
@@ -10627,7 +10629,7 @@ _NEON2SSE_INLINE uint16x4x4_t vld4_dup_u16(__transfersize(4) uint16_t const * pt
     return v;
 }
 
-uint32x2x4_t vld4_dup_u32(__transfersize(4) uint32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE uint32x2x4_t vld4_dup_u32(__transfersize(4) uint32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
 _NEON2SSE_INLINE uint32x2x4_t vld4_dup_u32(__transfersize(4) uint32_t const * ptr) // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
 {
     uint32x2x4_t v;
@@ -10644,7 +10646,7 @@ _NEON2SSE_INLINE uint32x2x4_t vld4_dup_u32(__transfersize(4) uint32_t const * pt
     return v;
 }
 
-uint64x1x4_t vld4_dup_u64(__transfersize(4) uint64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE uint64x1x4_t vld4_dup_u64(__transfersize(4) uint64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
 _NEON2SSE_INLINE uint64x1x4_t vld4_dup_u64(__transfersize(4) uint64_t const * ptr) // VLD1.64 {d0, d1, d2, d3}, [r0]
 {
     uint64x1x4_t v;
@@ -10655,22 +10657,22 @@ _NEON2SSE_INLINE uint64x1x4_t vld4_dup_u64(__transfersize(4) uint64_t const * pt
     return v;
 }
 
-int8x8x4_t vld4_dup_s8(__transfersize(4) int8_t const * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE int8x8x4_t vld4_dup_s8(__transfersize(4) int8_t const * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
 #define vld4_dup_s8(ptr) vld4_dup_u8((uint8_t*)ptr)
 
-int16x4x4_t vld4_dup_s16(__transfersize(4) int16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE int16x4x4_t vld4_dup_s16(__transfersize(4) int16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
 #define vld4_dup_s16(ptr) vld4_dup_u16((uint16_t*)ptr)
 
-int32x2x4_t vld4_dup_s32(__transfersize(4) int32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE int32x2x4_t vld4_dup_s32(__transfersize(4) int32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
 #define vld4_dup_s32(ptr) vld4_dup_u32((uint32_t*)ptr)
 
 int64x1x4_t vld4_dup_s64(__transfersize(4) int64_t const * ptr); // VLD1.64 {d0, d1, d2, d3}, [r0]
 #define vld4_dup_s64(ptr) vld4_dup_u64((uint64_t*)ptr)
 
-float16x4x4_t vld4_dup_f16(__transfersize(4) __fp16 const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE float16x4x4_t vld4_dup_f16(__transfersize(4) __fp16 const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers. See vld1q_f16 for example
 
-float32x2x4_t vld4_dup_f32(__transfersize(4) float32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE float32x2x4_t vld4_dup_f32(__transfersize(4) float32_t const * ptr); // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
 _NEON2SSE_INLINE float32x2x4_t vld4_dup_f32(__transfersize(4) float32_t const * ptr) // VLD4.32 {d0[], d1[], d2[], d3[]}, [r0]
 {
     float32x2x4_t v;
@@ -10682,10 +10684,10 @@ _NEON2SSE_INLINE float32x2x4_t vld4_dup_f32(__transfersize(4) float32_t const *
     return v;
 }
 
-poly8x8x4_t vld4_dup_p8(__transfersize(4) poly8_t const  * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE poly8x8x4_t vld4_dup_p8(__transfersize(4) poly8_t const  * ptr); // VLD4.8 {d0[], d1[], d2[], d3[]}, [r0]
 #define vld4_dup_p8 vld4_dup_u8
 
-poly16x4x4_t vld4_dup_p16(__transfersize(4) poly16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
+_NEON2SSESTORAGE poly16x4x4_t vld4_dup_p16(__transfersize(4) poly16_t const * ptr); // VLD4.16 {d0[], d1[], d2[], d3[]}, [r0]
 #define vld4_dup_p16 vld4_dup_u16
 
 
@@ -10755,7 +10757,7 @@ _NEON2SSE_INLINE float32x4x2_t vld2q_lane_f32_ptr(__transfersize(2) float32_t co
 //poly16x8x2_t vld2q_lane_p16(__transfersize(2) poly16_t const * ptr, poly16x8x2_t src,__constrange(0,7) int lane);// VLD2.16 {d0[0], d2[0]}, [r0]
 #define vld2q_lane_p16 vld2q_lane_u16
 
-uint8x8x2_t vld2_lane_u8(__transfersize(2) uint8_t const * ptr, uint8x8x2_t src, __constrange(0,7) int lane);// VLD2.8 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE uint8x8x2_t vld2_lane_u8(__transfersize(2) uint8_t const * ptr, uint8x8x2_t src, __constrange(0,7) int lane);// VLD2.8 {d0[0], d1[0]}, [r0]
 _NEON2SSE_INLINE uint8x8x2_t vld2_lane_u8(__transfersize(2) uint8_t const * ptr, uint8x8x2_t src, __constrange(0,7) int lane) // VLD2.8 {d0[0], d1[0]}, [r0]
 {
     uint8x8x2_t v;
@@ -10764,7 +10766,7 @@ _NEON2SSE_INLINE uint8x8x2_t vld2_lane_u8(__transfersize(2) uint8_t const * ptr,
     return v;
 }
 
-uint16x4x2_t vld2_lane_u16(__transfersize(2) uint16_t const * ptr, uint16x4x2_t src, __constrange(0,3)int lane);// VLD2.16 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE uint16x4x2_t vld2_lane_u16(__transfersize(2) uint16_t const * ptr, uint16x4x2_t src, __constrange(0,3)int lane);// VLD2.16 {d0[0], d1[0]}, [r0]
 _NEON2SSE_INLINE uint16x4x2_t vld2_lane_u16(__transfersize(2) uint16_t const * ptr, uint16x4x2_t src, __constrange(0,3) int lane)
 {
     uint16x4x2_t v;
@@ -10773,7 +10775,7 @@ _NEON2SSE_INLINE uint16x4x2_t vld2_lane_u16(__transfersize(2) uint16_t const * p
     return v;
 }
 
-uint32x2x2_t vld2_lane_u32(__transfersize(2) uint32_t const * ptr, uint32x2x2_t src, __constrange(0,1)int lane);// VLD2.32 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE uint32x2x2_t vld2_lane_u32(__transfersize(2) uint32_t const * ptr, uint32x2x2_t src, __constrange(0,1)int lane);// VLD2.32 {d0[0], d1[0]}, [r0]
 _NEON2SSE_INLINE uint32x2x2_t vld2_lane_u32(__transfersize(2) uint32_t const * ptr, uint32x2x2_t src, __constrange(0,1) int lane)
 {
     uint32x2x2_t v;
@@ -10782,19 +10784,19 @@ _NEON2SSE_INLINE uint32x2x2_t vld2_lane_u32(__transfersize(2) uint32_t const * p
     return v;
 }
 
-int8x8x2_t vld2_lane_s8(__transfersize(2) int8_t const * ptr, int8x8x2_t src, __constrange(0,7) int lane);// VLD2.8 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE int8x8x2_t vld2_lane_s8(__transfersize(2) int8_t const * ptr, int8x8x2_t src, __constrange(0,7) int lane);// VLD2.8 {d0[0], d1[0]}, [r0]
 #define vld2_lane_s8(ptr, src, lane)  vld2_lane_u8(( uint8_t*) ptr, src, lane)
 
-int16x4x2_t vld2_lane_s16(__transfersize(2) int16_t const * ptr, int16x4x2_t src, __constrange(0,3) int lane);// VLD2.16 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE int16x4x2_t vld2_lane_s16(__transfersize(2) int16_t const * ptr, int16x4x2_t src, __constrange(0,3) int lane);// VLD2.16 {d0[0], d1[0]}, [r0]
 #define vld2_lane_s16(ptr, src, lane) vld2_lane_u16(( uint16_t*) ptr, src, lane)
 
-int32x2x2_t vld2_lane_s32(__transfersize(2) int32_t const * ptr, int32x2x2_t src, __constrange(0,1) int lane);// VLD2.32 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE int32x2x2_t vld2_lane_s32(__transfersize(2) int32_t const * ptr, int32x2x2_t src, __constrange(0,1) int lane);// VLD2.32 {d0[0], d1[0]}, [r0]
 #define vld2_lane_s32(ptr, src, lane) vld2_lane_u32(( uint32_t*) ptr, src, lane)
 
 //float16x4x2_t vld2_lane_f16(__transfersize(2) __fp16 const * ptr, float16x4x2_t src, __constrange(0,3) int lane); // VLD2.16 {d0[0], d1[0]}, [r0]
 //current IA SIMD doesn't support float16
 
-float32x2x2_t vld2_lane_f32(__transfersize(2) float32_t const * ptr, float32x2x2_t src,__constrange(0,1) int lane); // VLD2.32 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE float32x2x2_t vld2_lane_f32(__transfersize(2) float32_t const * ptr, float32x2x2_t src,__constrange(0,1) int lane); // VLD2.32 {d0[0], d1[0]}, [r0]
 _NEON2SSE_INLINE float32x2x2_t vld2_lane_f32(__transfersize(2) float32_t const * ptr, float32x2x2_t  src,__constrange(0,1) int lane)
 {
     float32x2x2_t v;
@@ -10804,11 +10806,11 @@ _NEON2SSE_INLINE float32x2x2_t vld2_lane_f32(__transfersize(2) float32_t const *
 }
 
 //poly8x8x2_t vld2_lane_p8(__transfersize(2) poly8_t const * ptr, poly8x8x2_t src, __constrange(0,7) int lane);// VLD2.8 {d0[0], d1[0]}, [r0]
-poly8x8x2_t vld2_lane_p8_ptr(__transfersize(2) poly8_t const * ptr, poly8x8x2_t * src, __constrange(0,7) int lane); // VLD2.8 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE poly8x8x2_t vld2_lane_p8_ptr(__transfersize(2) poly8_t const * ptr, poly8x8x2_t * src, __constrange(0,7) int lane); // VLD2.8 {d0[0], d1[0]}, [r0]
 #define vld2_lane_p8 vld2_lane_u8
 
 //poly16x4x2_t vld2_lane_p16(__transfersize(2) poly16_t const * ptr, poly16x4x2_t src, __constrange(0,3)int lane);// VLD2.16 {d0[0], d1[0]}, [r0]
-poly16x4x2_t vld2_lane_p16_ptr(__transfersize(2) poly16_t const * ptr, poly16x4x2_t * src, __constrange(0,3) int lane); // VLD2.16 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE poly16x4x2_t vld2_lane_p16_ptr(__transfersize(2) poly16_t const * ptr, poly16x4x2_t * src, __constrange(0,3) int lane); // VLD2.16 {d0[0], d1[0]}, [r0]
 #define vld2_lane_p16 vld2_lane_u16
 
 //*********** Lane triplets **********************
@@ -10860,7 +10862,7 @@ _NEON2SSE_INLINE int32x4x3_t vld3q_lane_s32_ptr(__transfersize(3) int32_t const
 }
 #define vld3q_lane_s32(ptr, src, lane) vld3q_lane_s32_ptr(ptr, &src, lane)
 
-float16x8x3_t vld3q_lane_f16_ptr(__transfersize(3) __fp16 const * ptr, float16x8x3_t * src, __constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE float16x8x3_t vld3q_lane_f16_ptr(__transfersize(3) __fp16 const * ptr, float16x8x3_t * src, __constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
 //current IA SIMD doesn't support float16
 #define vld3q_lane_f16(ptr, src, lane) vld3q_lane_f16_ptr(ptr, &src, lane)
 
@@ -10876,10 +10878,10 @@ _NEON2SSE_INLINE float32x4x3_t vld3q_lane_f32_ptr(__transfersize(3) float32_t co
 }
 #define vld3q_lane_f32(ptr,src,lane) vld3q_lane_f32_ptr(ptr,&src,lane)
 
-poly16x8x3_t vld3q_lane_p16_ptr(__transfersize(3) poly16_t const * ptr, poly16x8x3_t * src,__constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
+_NEON2SSESTORAGE poly16x8x3_t vld3q_lane_p16_ptr(__transfersize(3) poly16_t const * ptr, poly16x8x3_t * src,__constrange(0,7) int lane); // VLD3.16 {d0[0], d2[0], d4[0]}, [r0]
 #define vld3q_lane_p16 vld3q_lane_u16
 
-uint8x8x3_t vld3_lane_u8(__transfersize(3) uint8_t const * ptr, uint8x8x3_t src, __constrange(0,7) int lane);// VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE uint8x8x3_t vld3_lane_u8(__transfersize(3) uint8_t const * ptr, uint8x8x3_t src, __constrange(0,7) int lane);// VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
 _NEON2SSE_INLINE uint8x8x3_t vld3_lane_u8(__transfersize(3) uint8_t const * ptr, uint8x8x3_t src, __constrange(0,7) int lane) // VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
 {
     uint8x8x3_t v;
@@ -10889,7 +10891,7 @@ _NEON2SSE_INLINE uint8x8x3_t vld3_lane_u8(__transfersize(3) uint8_t const * ptr,
     return v;
 }
 
-uint16x4x3_t vld3_lane_u16(__transfersize(3) uint16_t   const * ptr, uint16x4x3_t src, __constrange(0,3)int lane);// VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE uint16x4x3_t vld3_lane_u16(__transfersize(3) uint16_t   const * ptr, uint16x4x3_t src, __constrange(0,3)int lane);// VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
 _NEON2SSE_INLINE uint16x4x3_t vld3_lane_u16(__transfersize(3) uint16_t const * ptr, uint16x4x3_t src, __constrange(0,3) int lane) // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
 {
     uint16x4x3_t v;
@@ -10899,7 +10901,7 @@ _NEON2SSE_INLINE uint16x4x3_t vld3_lane_u16(__transfersize(3) uint16_t const * p
     return v;
 }
 
-uint32x2x3_t vld3_lane_u32(__transfersize(3) uint32_t const * ptr, uint32x2x3_t src, __constrange(0,1)int lane);// VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE uint32x2x3_t vld3_lane_u32(__transfersize(3) uint32_t const * ptr, uint32x2x3_t src, __constrange(0,1)int lane);// VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
 _NEON2SSE_INLINE uint32x2x3_t vld3_lane_u32(__transfersize(3) uint32_t const * ptr, uint32x2x3_t src, __constrange(0,1) int lane) // VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
 {
     //need to merge into 128 bit anyway
@@ -10910,19 +10912,19 @@ _NEON2SSE_INLINE uint32x2x3_t vld3_lane_u32(__transfersize(3) uint32_t const * p
     return v;
 }
 
-int8x8x3_t vld3_lane_s8(__transfersize(3) int8_t const * ptr, int8x8x3_t  src, __constrange(0,7) int lane); // VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE int8x8x3_t vld3_lane_s8(__transfersize(3) int8_t const * ptr, int8x8x3_t  src, __constrange(0,7) int lane); // VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
 #define vld3_lane_s8(ptr, src, lane)  vld3_lane_u8(( uint8_t*) ptr, src, lane)
 
-int16x4x3_t vld3_lane_s16(__transfersize(3) int16_t const * ptr, int16x4x3_t  src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE int16x4x3_t vld3_lane_s16(__transfersize(3) int16_t const * ptr, int16x4x3_t  src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
 #define vld3_lane_s16(ptr, src, lane)  vld3_lane_u16(( uint16_t*) ptr, src, lane)
 
-int32x2x3_t vld3_lane_s32(__transfersize(3) int32_t const * ptr, int32x2x3_t  src, __constrange(0,1) int lane); // VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE int32x2x3_t vld3_lane_s32(__transfersize(3) int32_t const * ptr, int32x2x3_t  src, __constrange(0,1) int lane); // VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
 #define vld3_lane_s32(ptr, src, lane)  vld3_lane_u32(( uint32_t*) ptr, src, lane)
 
-float16x4x3_t vld3_lane_f16_ptr(__transfersize(3) __fp16 const * ptr, float16x4x3_t * src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE float16x4x3_t vld3_lane_f16_ptr(__transfersize(3) __fp16 const * ptr, float16x4x3_t * src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
 //current IA SIMD doesn't support float16
 
-float32x2x3_t vld3_lane_f32(__transfersize(3) float32_t const * ptr, float32x2x3_t src,__constrange(0,1) int lane);// VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE float32x2x3_t vld3_lane_f32(__transfersize(3) float32_t const * ptr, float32x2x3_t src,__constrange(0,1) int lane);// VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
 _NEON2SSE_INLINE float32x2x3_t vld3_lane_f32(__transfersize(3) float32_t const * ptr, float32x2x3_t src,__constrange(0,1) int lane) // VLD3.32 {d0[0], d1[0], d2[0]}, [r0]
 {
     float32x2x3_t v;
@@ -10932,10 +10934,10 @@ _NEON2SSE_INLINE float32x2x3_t vld3_lane_f32(__transfersize(3) float32_t const *
     return v;
 }
 
-poly8x8x3_t vld3_lane_p8(__transfersize(3) poly8_t const * ptr, poly8x8x3_t src, __constrange(0,7) int lane); // VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE poly8x8x3_t vld3_lane_p8(__transfersize(3) poly8_t const * ptr, poly8x8x3_t src, __constrange(0,7) int lane); // VLD3.8 {d0[0], d1[0], d2[0]}, [r0]
 #define vld3_lane_p8 vld3_lane_u8
 
-poly16x4x3_t vld3_lane_p16(__transfersize(3) poly16_t const * ptr, poly16x4x3_t src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE poly16x4x3_t vld3_lane_p16(__transfersize(3) poly16_t const * ptr, poly16x4x3_t src, __constrange(0,3) int lane); // VLD3.16 {d0[0], d1[0], d2[0]}, [r0]
 #define vld3_lane_p16 vld3_lane_u16
 
 //******************* Lane Quadruples  load ***************************
@@ -10968,15 +10970,15 @@ _NEON2SSE_INLINE uint32x4x4_t vld4q_lane_u32_ptr(__transfersize(4) uint32_t cons
 #define vld4q_lane_u32(ptr, src, lane) vld4q_lane_u32_ptr(ptr, &src, lane)
 
 //int16x8x4_t vld4q_lane_s16(__transfersize(4) int16_t const * ptr, int16x8x4_t src, __constrange(0,7)int lane);// VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-int16x8x4_t vld4q_lane_s16_ptr(__transfersize(4) int16_t const * ptr, int16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE int16x8x4_t vld4q_lane_s16_ptr(__transfersize(4) int16_t const * ptr, int16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
 #define vld4q_lane_s16(ptr, src, lane) vld4q_lane_u16(( uint16_t*) ptr, src, lane)
 
 //int32x4x4_t vld4q_lane_s32(__transfersize(4) int32_t const * ptr, int32x4x4_t src, __constrange(0,3)int lane);// VLD4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-int32x4x4_t vld4q_lane_s32_ptr(__transfersize(4) int32_t const * ptr, int32x4x4_t * src, __constrange(0,3) int lane); // VLD4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE int32x4x4_t vld4q_lane_s32_ptr(__transfersize(4) int32_t const * ptr, int32x4x4_t * src, __constrange(0,3) int lane); // VLD4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
 #define vld4q_lane_s32(ptr, src, lane)  vld4q_lane_u32(( uint32_t*) ptr, src, lane)
 
 //float16x8x4_t vld4q_lane_f16(__transfersize(4) __fp16 const * ptr, float16x8x4_t src, __constrange(0,7)int lane);// VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-float16x8x4_t vld4q_lane_f16_ptr(__transfersize(4) __fp16 const * ptr, float16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE float16x8x4_t vld4q_lane_f16_ptr(__transfersize(4) __fp16 const * ptr, float16x8x4_t * src, __constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
 //current IA SIMD doesn't support float16
 
 //float32x4x4_t vld4q_lane_f32(__transfersize(4) float32_t const * ptr, float32x4x4_t src,__constrange(0,3) int lane)// VLD4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
@@ -10992,10 +10994,10 @@ _NEON2SSE_INLINE float32x4x4_t vld4q_lane_f32_ptr(__transfersize(4) float32_t co
 #define vld4q_lane_f32(ptr,val,lane) vld4q_lane_f32_ptr(ptr,&val,lane)
 
 //poly16x8x4_t vld4q_lane_p16(__transfersize(4) poly16_t const * ptr, poly16x8x4_t src,__constrange(0,7) int lane);// VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-poly16x8x4_t vld4q_lane_p16_ptr(__transfersize(4) poly16_t const * ptr, poly16x8x4_t * src,__constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
+_NEON2SSESTORAGE poly16x8x4_t vld4q_lane_p16_ptr(__transfersize(4) poly16_t const * ptr, poly16x8x4_t * src,__constrange(0,7) int lane); // VLD4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
 #define vld4q_lane_p16 vld4q_lane_u16
 
-uint8x8x4_t vld4_lane_u8(__transfersize(4) uint8_t const * ptr, uint8x8x4_t src, __constrange(0,7) int lane);// VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE uint8x8x4_t vld4_lane_u8(__transfersize(4) uint8_t const * ptr, uint8x8x4_t src, __constrange(0,7) int lane);// VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 _NEON2SSE_INLINE uint8x8x4_t vld4_lane_u8(__transfersize(4) uint8_t const * ptr, uint8x8x4_t src, __constrange(0,7) int lane)
 {
     uint8x8x4_t v;
@@ -11006,7 +11008,7 @@ _NEON2SSE_INLINE uint8x8x4_t vld4_lane_u8(__transfersize(4) uint8_t const * ptr,
     return v;
 }
 
-uint16x4x4_t vld4_lane_u16(__transfersize(4) uint16_t const * ptr, uint16x4x4_t src, __constrange(0,3)int lane);// VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE uint16x4x4_t vld4_lane_u16(__transfersize(4) uint16_t const * ptr, uint16x4x4_t src, __constrange(0,3)int lane);// VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 _NEON2SSE_INLINE uint16x4x4_t vld4_lane_u16(__transfersize(4) uint16_t const * ptr, uint16x4x4_t src, __constrange(0,3) int lane)
 {
     uint16x4x4_t v;
@@ -11017,7 +11019,7 @@ _NEON2SSE_INLINE uint16x4x4_t vld4_lane_u16(__transfersize(4) uint16_t const * p
     return v;
 }
 
-uint32x2x4_t vld4_lane_u32(__transfersize(4) uint32_t const * ptr, uint32x2x4_t src, __constrange(0,1)int lane);// VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE uint32x2x4_t vld4_lane_u32(__transfersize(4) uint32_t const * ptr, uint32x2x4_t src, __constrange(0,1)int lane);// VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 _NEON2SSE_INLINE uint32x2x4_t vld4_lane_u32(__transfersize(4) uint32_t const * ptr, uint32x2x4_t src, __constrange(0,1) int lane)
 {
     uint32x2x4_t v;
@@ -11028,20 +11030,20 @@ _NEON2SSE_INLINE uint32x2x4_t vld4_lane_u32(__transfersize(4) uint32_t const * p
     return v;
 }
 
-int8x8x4_t vld4_lane_s8(__transfersize(4) int8_t const * ptr, int8x8x4_t src, __constrange(0,7) int lane);// VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE int8x8x4_t vld4_lane_s8(__transfersize(4) int8_t const * ptr, int8x8x4_t src, __constrange(0,7) int lane);// VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 #define vld4_lane_s8(ptr,src,lane) vld4_lane_u8((uint8_t*)ptr,src,lane)
 
-int16x4x4_t vld4_lane_s16(__transfersize(4) int16_t const * ptr, int16x4x4_t src, __constrange(0,3) int lane);// VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE int16x4x4_t vld4_lane_s16(__transfersize(4) int16_t const * ptr, int16x4x4_t src, __constrange(0,3) int lane);// VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 #define vld4_lane_s16(ptr,src,lane) vld4_lane_u16((uint16_t*)ptr,src,lane)
 
-int32x2x4_t vld4_lane_s32(__transfersize(4) int32_t const * ptr, int32x2x4_t src, __constrange(0,1) int lane);// VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE int32x2x4_t vld4_lane_s32(__transfersize(4) int32_t const * ptr, int32x2x4_t src, __constrange(0,1) int lane);// VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 #define vld4_lane_s32(ptr,src,lane) vld4_lane_u32((uint32_t*)ptr,src,lane)
 
 //float16x4x4_t vld4_lane_f16(__transfersize(4) __fp16 const * ptr, float16x4x4_t src, __constrange(0,3)int lane);// VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
-float16x4x4_t vld4_lane_f16_ptr(__transfersize(4) __fp16 const * ptr, float16x4x4_t * src, __constrange(0,3) int lane);
+_NEON2SSESTORAGE float16x4x4_t vld4_lane_f16_ptr(__transfersize(4) __fp16 const * ptr, float16x4x4_t * src, __constrange(0,3) int lane);
 //current IA SIMD doesn't support float16
 
-float32x2x4_t vld4_lane_f32(__transfersize(4) float32_t const * ptr, float32x2x4_t src,__constrange(0,1) int lane);// VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE float32x2x4_t vld4_lane_f32(__transfersize(4) float32_t const * ptr, float32x2x4_t src,__constrange(0,1) int lane);// VLD4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 _NEON2SSE_INLINE float32x2x4_t vld4_lane_f32(__transfersize(4) float32_t const * ptr, float32x2x4_t src,__constrange(0,1) int lane)
 {
     //serial solution may be faster
@@ -11053,10 +11055,10 @@ _NEON2SSE_INLINE float32x2x4_t vld4_lane_f32(__transfersize(4) float32_t const *
     return v;
 }
 
-poly8x8x4_t vld4_lane_p8(__transfersize(4) poly8_t const * ptr, poly8x8x4_t src, __constrange(0,7) int lane);// VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE poly8x8x4_t vld4_lane_p8(__transfersize(4) poly8_t const * ptr, poly8x8x4_t src, __constrange(0,7) int lane);// VLD4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 #define vld4_lane_p8 vld4_lane_u8
 
-poly16x4x4_t vld4_lane_p16(__transfersize(4) poly16_t const * ptr, poly16x4x4_t src, __constrange(0,3)int lane);// VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE poly16x4x4_t vld4_lane_p16(__transfersize(4) poly16_t const * ptr, poly16x4x4_t src, __constrange(0,3)int lane);// VLD4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 #define vld4_lane_p16 vld4_lane_u16
 
 //******************* Store duplets *********************************************
@@ -11095,19 +11097,19 @@ _NEON2SSE_INLINE void vst2q_u32_ptr(__transfersize(8) uint32_t* ptr, uint32x4x2_
 #define vst2q_u32(ptr, val) vst2q_u32_ptr(ptr, &val)
 
 //void vst2q_s8(__transfersize(32) int8_t * ptr, int8x16x2_t val); // VST2.8 {d0, d2}, [r0]
-void vst2q_s8_ptr(__transfersize(32) int8_t * ptr, int8x16x2_t * val);
+_NEON2SSESTORAGE void vst2q_s8_ptr(__transfersize(32) int8_t * ptr, int8x16x2_t * val);
 #define vst2q_s8(ptr, val) vst2q_u8((uint8_t*)(ptr), val)
 
 //void vst2q_s16(__transfersize(16) int16_t * ptr, int16x8x2_t val);// VST2.16 {d0, d2}, [r0]
-void vst2q_s16_ptr(__transfersize(16) int16_t * ptr, int16x8x2_t * val);
+_NEON2SSESTORAGE void vst2q_s16_ptr(__transfersize(16) int16_t * ptr, int16x8x2_t * val);
 #define vst2q_s16(ptr, val) vst2q_u16((uint16_t*)(ptr), val)
 
 //void vst2q_s32(__transfersize(8) int32_t * ptr, int32x4x2_t val);// VST2.32 {d0, d2}, [r0]
-void vst2q_s32_ptr(__transfersize(8) int32_t * ptr, int32x4x2_t * val);
+_NEON2SSESTORAGE void vst2q_s32_ptr(__transfersize(8) int32_t * ptr, int32x4x2_t * val);
 #define vst2q_s32(ptr, val)  vst2q_u32((uint32_t*)(ptr), val)
 
 //void vst2q_f16(__transfersize(16) __fp16 * ptr, float16x8x2_t val);// VST2.16 {d0, d2}, [r0]
-void vst2q_f16_ptr(__transfersize(16) __fp16 * ptr, float16x8x2_t * val);
+_NEON2SSESTORAGE void vst2q_f16_ptr(__transfersize(16) __fp16 * ptr, float16x8x2_t * val);
 // IA32 SIMD doesn't work with 16bit floats currently
 
 //void vst2q_f32(__transfersize(8) float32_t * ptr, float32x4x2_t val)// VST2.32 {d0, d2}, [r0]
@@ -11122,14 +11124,14 @@ _NEON2SSE_INLINE void vst2q_f32_ptr(__transfersize(8) float32_t* ptr, float32x4x
 #define vst2q_f32(ptr, val) vst2q_f32_ptr(ptr, &val)
 
 //void vst2q_p8(__transfersize(32) poly8_t * ptr, poly8x16x2_t val);// VST2.8 {d0, d2}, [r0]
-void vst2q_p8_ptr(__transfersize(32) poly8_t * ptr, poly8x16x2_t * val);
+_NEON2SSESTORAGE void vst2q_p8_ptr(__transfersize(32) poly8_t * ptr, poly8x16x2_t * val);
 #define vst2q_p8 vst2q_u8
 
 //void vst2q_p16(__transfersize(16) poly16_t * ptr, poly16x8x2_t val);// VST2.16 {d0, d2}, [r0]
-void vst2q_p16_ptr(__transfersize(16) poly16_t * ptr, poly16x8x2_t * val);
+_NEON2SSESTORAGE void vst2q_p16_ptr(__transfersize(16) poly16_t * ptr, poly16x8x2_t * val);
 #define vst2q_p16 vst2q_u16
 
-void vst2_u8(__transfersize(16) uint8_t * ptr, uint8x8x2_t val);// VST2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_u8(__transfersize(16) uint8_t * ptr, uint8x8x2_t val);// VST2.8 {d0, d1}, [r0]
 _NEON2SSE_INLINE void vst2_u8(__transfersize(16) uint8_t * ptr, uint8x8x2_t val)
 {
     __m128i v0;
@@ -11137,7 +11139,7 @@ _NEON2SSE_INLINE void vst2_u8(__transfersize(16) uint8_t * ptr, uint8x8x2_t val)
     vst1q_u8 (ptr, v0);
 }
 
-void vst2_u16(__transfersize(8) uint16_t * ptr, uint16x4x2_t val);// VST2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_u16(__transfersize(8) uint16_t * ptr, uint16x4x2_t val);// VST2.16 {d0, d1}, [r0]
 _NEON2SSE_INLINE void vst2_u16(__transfersize(8) uint16_t * ptr, uint16x4x2_t val)
 {
     __m128i v0;
@@ -11145,7 +11147,7 @@ _NEON2SSE_INLINE void vst2_u16(__transfersize(8) uint16_t * ptr, uint16x4x2_t va
     vst1q_u16 (ptr, v0);
 }
 
-void vst2_u32(__transfersize(4) uint32_t * ptr, uint32x2x2_t val);// VST2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_u32(__transfersize(4) uint32_t * ptr, uint32x2x2_t val);// VST2.32 {d0, d1}, [r0]
 _NEON2SSE_INLINE void vst2_u32(__transfersize(4) uint32_t * ptr, uint32x2x2_t val)
 {
     __m128i v0;
@@ -11153,29 +11155,29 @@ _NEON2SSE_INLINE void vst2_u32(__transfersize(4) uint32_t * ptr, uint32x2x2_t va
     vst1q_u32 (ptr, v0);
 }
 
-void vst2_u64(__transfersize(2) uint64_t * ptr, uint64x1x2_t val);// VST1.64 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_u64(__transfersize(2) uint64_t * ptr, uint64x1x2_t val);// VST1.64 {d0, d1}, [r0]
 _NEON2SSE_INLINE void vst2_u64(__transfersize(2) uint64_t * ptr, uint64x1x2_t val)
 {
     *(ptr) = val.val[0].m64_u64[0];
     *(ptr + 1) = val.val[1].m64_u64[0];
 }
 
-void vst2_s8(__transfersize(16) int8_t * ptr, int8x8x2_t val);// VST2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_s8(__transfersize(16) int8_t * ptr, int8x8x2_t val);// VST2.8 {d0, d1}, [r0]
 #define vst2_s8(ptr, val) vst2_u8((uint8_t*) ptr, val)
 
-void vst2_s16(__transfersize(8) int16_t * ptr, int16x4x2_t val); // VST2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_s16(__transfersize(8) int16_t * ptr, int16x4x2_t val); // VST2.16 {d0, d1}, [r0]
 #define vst2_s16(ptr,val) vst2_u16((uint16_t*) ptr, val)
 
-void vst2_s32(__transfersize(4) int32_t * ptr, int32x2x2_t val); // VST2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_s32(__transfersize(4) int32_t * ptr, int32x2x2_t val); // VST2.32 {d0, d1}, [r0]
 #define vst2_s32(ptr,val) vst2_u32((uint32_t*) ptr, val)
 
-void vst2_s64(__transfersize(2) int64_t * ptr, int64x1x2_t val);
+_NEON2SSESTORAGE void vst2_s64(__transfersize(2) int64_t * ptr, int64x1x2_t val);
 #define vst2_s64(ptr,val) vst2_u64((uint64_t*) ptr,val)
 
 //void vst2_f16(__transfersize(8) __fp16 * ptr, float16x4x2_t val); // VST2.16 {d0, d1}, [r0]
 //current IA SIMD doesn't support float16
 
-void vst2_f32(__transfersize(4) float32_t * ptr, float32x2x2_t val); // VST2.32 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_f32(__transfersize(4) float32_t * ptr, float32x2x2_t val); // VST2.32 {d0, d1}, [r0]
 _NEON2SSE_INLINE void vst2_f32(__transfersize(4) float32_t* ptr, float32x2x2_t val)
 {
     *(ptr) =   val.val[0].m64_f32[0];
@@ -11184,10 +11186,10 @@ _NEON2SSE_INLINE void vst2_f32(__transfersize(4) float32_t* ptr, float32x2x2_t v
     *(ptr + 3) = val.val[1].m64_f32[1];
 }
 
-void vst2_p8(__transfersize(16) poly8_t * ptr, poly8x8x2_t  val); // VST2.8 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_p8(__transfersize(16) poly8_t * ptr, poly8x8x2_t  val); // VST2.8 {d0, d1}, [r0]
 #define vst2_p8 vst2_u8
 
-void vst2_p16(__transfersize(8) poly16_t * ptr, poly16x4x2_t  val); // VST2.16 {d0, d1}, [r0]
+_NEON2SSESTORAGE void vst2_p16(__transfersize(8) poly16_t * ptr, poly16x4x2_t  val); // VST2.16 {d0, d1}, [r0]
 #define vst2_p16 vst2_u16
 
 //******************** Triplets store  *****************************************
@@ -11197,12 +11199,12 @@ _NEON2SSE_INLINE void vst3q_u8_ptr(__transfersize(48) uint8_t * ptr, uint8x16x3_
 {
     uint8x16x3_t v;
     __m128i v0,v1,v2, cff, bldmask;
-    _NEON2SSE_ALIGN_16 uint8_t mask0[16]   = {0, 1, 0xff, 2, 3,0xff, 4, 5,0xff, 6,7,0xff, 8,9,0xff, 10};
-    _NEON2SSE_ALIGN_16 uint8_t mask1[16]   = {0, 0xff, 1, 2, 0xff, 3, 4, 0xff, 5, 6, 0xff, 7,8,0xff, 9,10};
-    _NEON2SSE_ALIGN_16 uint8_t mask2[16] =    {0xff, 6, 7, 0xff, 8, 9,0xff, 10, 11,0xff, 12,13,0xff, 14,15,0xff};
-    _NEON2SSE_ALIGN_16 uint8_t mask2lo[16] = {0xff,0xff, 0, 0xff,0xff, 1, 0xff,0xff, 2, 0xff,0xff, 3, 0xff,0xff, 4, 0xff};
-    _NEON2SSE_ALIGN_16 uint8_t mask2med[16] = {0xff, 5, 0xff, 0xff, 6, 0xff,0xff, 7, 0xff,0xff, 8, 0xff,0xff, 9, 0xff, 0xff};
-    _NEON2SSE_ALIGN_16 uint8_t mask2hi[16] = {10, 0xff,0xff, 11, 0xff,0xff, 12, 0xff,0xff, 13, 0xff,0xff, 14, 0xff, 0xff, 15};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask0[16]   = {0, 1, 0xff, 2, 3,0xff, 4, 5,0xff, 6,7,0xff, 8,9,0xff, 10};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask1[16]   = {0, 0xff, 1, 2, 0xff, 3, 4, 0xff, 5, 6, 0xff, 7,8,0xff, 9,10};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask2[16] =    {0xff, 6, 7, 0xff, 8, 9,0xff, 10, 11,0xff, 12,13,0xff, 14,15,0xff};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask2lo[16] = {0xff,0xff, 0, 0xff,0xff, 1, 0xff,0xff, 2, 0xff,0xff, 3, 0xff,0xff, 4, 0xff};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask2med[16] = {0xff, 5, 0xff, 0xff, 6, 0xff,0xff, 7, 0xff,0xff, 8, 0xff,0xff, 9, 0xff, 0xff};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask2hi[16] = {10, 0xff,0xff, 11, 0xff,0xff, 12, 0xff,0xff, 13, 0xff,0xff, 14, 0xff, 0xff, 15};
 
     v0 =  _mm_unpacklo_epi8(val->val[0], val->val[1]); //0,1, 3,4, 6,7, 9,10, 12,13, 15,16, 18,19, 21,22
     v2 =  _mm_unpackhi_epi8(val->val[0], val->val[1]); //24,25,  27,28, 30,31, 33,34, 36,37, 39,40, 42,43, 45,46
@@ -11231,12 +11233,12 @@ _NEON2SSE_INLINE void vst3q_u16_ptr(__transfersize(24) uint16_t * ptr, uint16x8x
 {
     uint16x8x3_t v;
     __m128i v0,v1,v2, cff, bldmask;
-    _NEON2SSE_ALIGN_16 uint8_t mask0[16]   = {0,1, 2,3, 0xff,0xff, 4,5, 6,7,0xff,0xff, 8,9,10,11};
-    _NEON2SSE_ALIGN_16 uint8_t mask1[16]   = {0xff, 0xff, 0,1, 2,3, 0xff,0xff, 4,5, 6,7, 0xff,0xff, 8,9};
-    _NEON2SSE_ALIGN_16 uint8_t mask2[16] =    {6,7,0xff,0xff, 8,9,10,11, 0xff, 0xff, 12,13,14,15, 0xff, 0xff};
-    _NEON2SSE_ALIGN_16 uint8_t mask2lo[16] = {0xff,0xff, 0xff,0xff, 0,1, 0xff,0xff, 0xff,0xff, 2,3, 0xff,0xff, 0xff,0xff};
-    _NEON2SSE_ALIGN_16 uint8_t mask2med[16] = {4,5, 0xff,0xff,0xff,0xff, 6,7, 0xff, 0xff,0xff,0xff, 8,9, 0xff, 0xff};
-    _NEON2SSE_ALIGN_16 uint8_t mask2hi[16] = {0xff, 0xff, 10,11, 0xff, 0xff, 0xff, 0xff, 12,13, 0xff, 0xff, 0xff, 0xff,14,15};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask0[16]   = {0,1, 2,3, 0xff,0xff, 4,5, 6,7,0xff,0xff, 8,9,10,11};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask1[16]   = {0xff, 0xff, 0,1, 2,3, 0xff,0xff, 4,5, 6,7, 0xff,0xff, 8,9};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask2[16] =    {6,7,0xff,0xff, 8,9,10,11, 0xff, 0xff, 12,13,14,15, 0xff, 0xff};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask2lo[16] = {0xff,0xff, 0xff,0xff, 0,1, 0xff,0xff, 0xff,0xff, 2,3, 0xff,0xff, 0xff,0xff};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask2med[16] = {4,5, 0xff,0xff,0xff,0xff, 6,7, 0xff, 0xff,0xff,0xff, 8,9, 0xff, 0xff};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask2hi[16] = {0xff, 0xff, 10,11, 0xff, 0xff, 0xff, 0xff, 12,13, 0xff, 0xff, 0xff, 0xff,14,15};
 
     v0 =  _mm_unpacklo_epi16(val->val[0], val->val[1]); //0,1, 3,4, 6,7, 9,10
     v2 =  _mm_unpackhi_epi16(val->val[0], val->val[1]); //12,13, 15,16, 18,19, 21,22,
@@ -11282,19 +11284,19 @@ _NEON2SSE_INLINE void vst3q_u32_ptr(__transfersize(12) uint32_t * ptr, uint32x4x
 #define vst3q_u32(ptr, val) vst3q_u32_ptr(ptr, &val)
 
 //void vst3q_s8(__transfersize(48) int8_t * ptr, int8x16x3_t val);
-void vst3q_s8_ptr(__transfersize(48) int8_t * ptr, int8x16x3_t * val);
+_NEON2SSESTORAGE void vst3q_s8_ptr(__transfersize(48) int8_t * ptr, int8x16x3_t * val);
 #define vst3q_s8(ptr, val) vst3q_u8((uint8_t*)(ptr), val)
 
 //void vst3q_s16(__transfersize(24) int16_t * ptr, int16x8x3_t val);
-void vst3q_s16_ptr(__transfersize(24) int16_t * ptr, int16x8x3_t * val);
+_NEON2SSESTORAGE void vst3q_s16_ptr(__transfersize(24) int16_t * ptr, int16x8x3_t * val);
 #define vst3q_s16(ptr, val) vst3q_u16((uint16_t*)(ptr), val)
 
 //void vst3q_s32(__transfersize(12) int32_t * ptr, int32x4x3_t val);
-void vst3q_s32_ptr(__transfersize(12) int32_t * ptr, int32x4x3_t * val);
+_NEON2SSESTORAGE void vst3q_s32_ptr(__transfersize(12) int32_t * ptr, int32x4x3_t * val);
 #define vst3q_s32(ptr, val)  vst3q_u32((uint32_t*)(ptr), val)
 
 //void vst3q_f16(__transfersize(24) __fp16 * ptr, float16x8x3_t val);// VST3.16 {d0, d2, d4}, [r0]
-void vst3q_f16_ptr(__transfersize(24) __fp16 * ptr, float16x8x3_t * val);
+_NEON2SSESTORAGE void vst3q_f16_ptr(__transfersize(24) __fp16 * ptr, float16x8x3_t * val);
 // IA32 SIMD doesn't work with 16bit floats currently
 
 //void vst3q_f32(__transfersize(12) float32_t * ptr, float32x4x3_t val)// VST3.32 {d0, d2, d4}, [r0]
@@ -11318,21 +11320,21 @@ _NEON2SSE_INLINE void vst3q_f32_ptr(__transfersize(12) float32_t * ptr, float32x
 #define vst3q_f32(ptr, val) vst3q_f32_ptr(ptr, &val)
 
 //void vst3q_p8(__transfersize(48) poly8_t * ptr, poly8x16x3_t val);// VST3.8 {d0, d2, d4}, [r0]
-void vst3q_p8_ptr(__transfersize(48) poly8_t * ptr, poly8x16x3_t * val);
+_NEON2SSESTORAGE void vst3q_p8_ptr(__transfersize(48) poly8_t * ptr, poly8x16x3_t * val);
 #define vst3q_p8 vst3q_u8
 
 //void vst3q_p16(__transfersize(24) poly16_t * ptr, poly16x8x3_t val);// VST3.16 {d0, d2, d4}, [r0]
-void vst3q_p16_ptr(__transfersize(24) poly16_t * ptr, poly16x8x3_t * val);
+_NEON2SSESTORAGE void vst3q_p16_ptr(__transfersize(24) poly16_t * ptr, poly16x8x3_t * val);
 #define vst3q_p16 vst3q_u16
 
-void vst3_u8(__transfersize(24) uint8_t * ptr, uint8x8x3_t val);// VST3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_u8(__transfersize(24) uint8_t * ptr, uint8x8x3_t val);// VST3.8 {d0, d1, d2}, [r0]
 _NEON2SSE_INLINE void vst3_u8(__transfersize(24) uint8_t * ptr, uint8x8x3_t val)
 {
     __m128i tmp, sh0, sh1, val0, val2;
-    _NEON2SSE_ALIGN_16 int8_t mask0[16] = { 0, 8, 16, 1, 9, 17, 2, 10, 18, 3, 11, 19, 4, 12, 20, 5};
-    _NEON2SSE_ALIGN_16 int8_t mask1[16] = {13, 21, 6, 14, 22, 7, 15, 23, 0,0,0,0,0,0,0,0};
-    _NEON2SSE_ALIGN_16 uint8_t mask0_sel[16] = {0, 0, 0xff, 0, 0, 0xff, 0, 0, 0xff, 0, 0, 0xff, 0, 0, 0xff, 0};
-    _NEON2SSE_ALIGN_16 uint8_t mask1_sel[16] = {0, 0xff, 0, 0, 0xff, 0, 0, 0xff, 0,0,0,0,0,0,0,0};
+    _NEON2SSE_ALIGN_16 static const int8_t mask0[16] = { 0, 8, 16, 1, 9, 17, 2, 10, 18, 3, 11, 19, 4, 12, 20, 5};
+    _NEON2SSE_ALIGN_16 static const int8_t mask1[16] = {13, 21, 6, 14, 22, 7, 15, 23, 0,0,0,0,0,0,0,0};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask0_sel[16] = {0, 0, 0xff, 0, 0, 0xff, 0, 0, 0xff, 0, 0, 0xff, 0, 0, 0xff, 0};
+    _NEON2SSE_ALIGN_16 static const uint8_t mask1_sel[16] = {0, 0xff, 0, 0, 0xff, 0, 0, 0xff, 0,0,0,0,0,0,0,0};
     tmp = _mm_unpacklo_epi64(_pM128i(val.val[0]), _pM128i(val.val[1]) );
     sh0 =  _mm_shuffle_epi8(tmp, *(__m128i*)mask0); //for bi>15 bi is wrapped (bi-=15)
     val2 = _pM128i(val.val[2]);
@@ -11345,14 +11347,14 @@ _NEON2SSE_INLINE void vst3_u8(__transfersize(24) uint8_t * ptr, uint8x8x3_t val)
     _M64((*(__m64_128*)(ptr + 16)),  val2); //need it to fit into *ptr memory
 }
 
-void vst3_u16(__transfersize(12) uint16_t * ptr, uint16x4x3_t val);// VST3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_u16(__transfersize(12) uint16_t * ptr, uint16x4x3_t val);// VST3.16 {d0, d1, d2}, [r0]
 _NEON2SSE_INLINE void vst3_u16(__transfersize(12) uint16_t * ptr, uint16x4x3_t val)
 {
     __m128i tmp, val0, val1, val2;
-    _NEON2SSE_ALIGN_16 int8_t mask0[16] = {0,1, 8,9, 16,17, 2,3, 10,11, 18,19, 4,5, 12,13};
-    _NEON2SSE_ALIGN_16 int8_t mask1[16] = {20,21, 6,7, 14,15, 22,23,   0,0,0,0,0,0,0,0};
-    _NEON2SSE_ALIGN_16 uint16_t mask0f[8] = {0xffff, 0xffff, 0, 0xffff, 0xffff, 0, 0xffff, 0xffff}; //if all ones we take the result from v.val[0]  otherwise from v.val[1]
-    _NEON2SSE_ALIGN_16 uint16_t mask1f[8] = {0xffff, 0, 0, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff}; //if all ones we take the result from v.val[1]  otherwise from v.val[0]
+    _NEON2SSE_ALIGN_16 static const int8_t mask0[16] = {0,1, 8,9, 16,17, 2,3, 10,11, 18,19, 4,5, 12,13};
+    _NEON2SSE_ALIGN_16 static const int8_t mask1[16] = {20,21, 6,7, 14,15, 22,23,   0,0,0,0,0,0,0,0};
+    _NEON2SSE_ALIGN_16 static const uint16_t mask0f[8] = {0xffff, 0xffff, 0, 0xffff, 0xffff, 0, 0xffff, 0xffff}; //if all ones we take the result from v.val[0]  otherwise from v.val[1]
+    _NEON2SSE_ALIGN_16 static const uint16_t mask1f[8] = {0xffff, 0, 0, 0xffff, 0xffff, 0xffff, 0xffff, 0xffff}; //if all ones we take the result from v.val[1]  otherwise from v.val[0]
     tmp = _mm_unpacklo_epi64(_pM128i(val.val[0]), _pM128i(val.val[1]));
     val0 = _mm_shuffle_epi8(tmp, *(__m128i*)mask0);
     val2 = _pM128i(val.val[2]);
@@ -11365,7 +11367,7 @@ _NEON2SSE_INLINE void vst3_u16(__transfersize(12) uint16_t * ptr, uint16x4x3_t v
     _M64((*(__m64_128*)(ptr + 8)),  val1); //need it to fit into *ptr memory
 }
 
-void vst3_u32(__transfersize(6) uint32_t * ptr, uint32x2x3_t val);// VST3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_u32(__transfersize(6) uint32_t * ptr, uint32x2x3_t val);// VST3.32 {d0, d1, d2}, [r0]
 _NEON2SSE_INLINE void vst3_u32(__transfersize(6) uint32_t * ptr, uint32x2x3_t val)
 {
     //val.val[0]:0,3,val.val[1]:1,4; val.val[2]:2,5,x,x;
@@ -11379,7 +11381,7 @@ _NEON2SSE_INLINE void vst3_u32(__transfersize(6) uint32_t * ptr, uint32x2x3_t va
     vst1q_u32(ptr, val0); //store as 128 bit structure
 }
 
-void vst3_u64(__transfersize(3) uint64_t * ptr, uint64x1x3_t val);// VST1.64 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_u64(__transfersize(3) uint64_t * ptr, uint64x1x3_t val);// VST1.64 {d0, d1, d2}, [r0]
 _NEON2SSE_INLINE void vst3_u64(__transfersize(3) uint64_t * ptr, uint64x1x3_t val)
 {
     *(ptr) = val.val[0].m64_u64[0];
@@ -11387,23 +11389,23 @@ _NEON2SSE_INLINE void vst3_u64(__transfersize(3) uint64_t * ptr, uint64x1x3_t va
     *(ptr + 2) = val.val[2].m64_u64[0];
 }
 
-void vst3_s8(__transfersize(24) int8_t * ptr, int8x8x3_t val);  // VST3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_s8(__transfersize(24) int8_t * ptr, int8x8x3_t val);  // VST3.8 {d0, d1, d2}, [r0]
 #define vst3_s8(ptr, val) vst3_u8((uint8_t*)ptr, val)
 
-void vst3_s16(__transfersize(12) int16_t * ptr, int16x4x3_t val);  // VST3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_s16(__transfersize(12) int16_t * ptr, int16x4x3_t val);  // VST3.16 {d0, d1, d2}, [r0]
 #define vst3_s16(ptr, val) vst3_u16((uint16_t*)ptr, val)
 
-void vst3_s32(__transfersize(6) int32_t * ptr, int32x2x3_t val); // VST3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_s32(__transfersize(6) int32_t * ptr, int32x2x3_t val); // VST3.32 {d0, d1, d2}, [r0]
 #define vst3_s32(ptr, val) vst3_u32((uint32_t*)ptr, val)
 
-void vst3_s64(__transfersize(3) int64_t * ptr, int64x1x3_t val); // VST1.64 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_s64(__transfersize(3) int64_t * ptr, int64x1x3_t val); // VST1.64 {d0, d1, d2}, [r0]
 #define vst3_s64(ptr, val) vst3_u64((uint64_t*)ptr, val)
 
 //void vst3_f16(__transfersize(12) __fp16 * ptr, float16x4x3_t val);// VST3.16 {d0, d1, d2}, [r0]
-void vst3_f16_ptr(__transfersize(12) __fp16 * ptr, float16x4x3_t * val); // VST3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_f16_ptr(__transfersize(12) __fp16 * ptr, float16x4x3_t * val); // VST3.16 {d0, d1, d2}, [r0]
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers. See vld1q_f16 for example
 
-void vst3_f32(__transfersize(6) float32_t * ptr, float32x2x3_t val);// VST3.32 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_f32(__transfersize(6) float32_t * ptr, float32x2x3_t val);// VST3.32 {d0, d1, d2}, [r0]
 _NEON2SSE_INLINE void vst3_f32(__transfersize(6) float32_t * ptr, float32x2x3_t val)
 {
     //val->val[0]:0,3,val->val[1]:1,4; val->val[2]:2,5,x,x;   -> 0,2, 4,1, 3,5
@@ -11415,10 +11417,10 @@ _NEON2SSE_INLINE void vst3_f32(__transfersize(6) float32_t * ptr, float32x2x3_t
     *(ptr + 5) = val.val[2].m64_f32[1];
 }
 
-void vst3_p8(__transfersize(24) poly8_t * ptr, poly8x8x3_t val);// VST3.8 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_p8(__transfersize(24) poly8_t * ptr, poly8x8x3_t val);// VST3.8 {d0, d1, d2}, [r0]
 #define vst3_p8 vst3_u8
 
-void vst3_p16(__transfersize(12) poly16_t * ptr, poly16x4x3_t val);// VST3.16 {d0, d1, d2}, [r0]
+_NEON2SSESTORAGE void vst3_p16(__transfersize(12) poly16_t * ptr, poly16x4x3_t val);// VST3.16 {d0, d1, d2}, [r0]
 #define vst3_p16 vst3_u16
 
 //***************  Quadruples store ********************************
@@ -11483,19 +11485,19 @@ _NEON2SSE_INLINE void vst4q_u32_ptr(__transfersize(16) uint32_t * ptr, uint32x4x
 #define vst4q_u32(ptr, val) vst4q_u32_ptr(ptr, &val)
 
 //void vst4q_s8(__transfersize(64) int8_t * ptr, int8x16x4_t val);
-void vst4q_s8_ptr(__transfersize(64) int8_t * ptr, int8x16x4_t * val);
+_NEON2SSESTORAGE void vst4q_s8_ptr(__transfersize(64) int8_t * ptr, int8x16x4_t * val);
 #define vst4q_s8(ptr, val) vst4q_u8((uint8_t*)(ptr), val)
 
 //void vst4q_s16(__transfersize(32) int16_t * ptr, int16x8x4_t val);
-void vst4q_s16_ptr(__transfersize(32) int16_t * ptr, int16x8x4_t * val);
+_NEON2SSESTORAGE void vst4q_s16_ptr(__transfersize(32) int16_t * ptr, int16x8x4_t * val);
 #define vst4q_s16(ptr, val) vst4q_u16((uint16_t*)(ptr), val)
 
 //void vst4q_s32(__transfersize(16) int32_t * ptr, int32x4x4_t val);
-void vst4q_s32_ptr(__transfersize(16) int32_t * ptr, int32x4x4_t * val);
+_NEON2SSESTORAGE void vst4q_s32_ptr(__transfersize(16) int32_t * ptr, int32x4x4_t * val);
 #define vst4q_s32(ptr, val) vst4q_u32((uint32_t*)(ptr), val)
 
 //void vst4q_f16(__transfersize(32) __fp16 * ptr, float16x8x4_t val);// VST4.16 {d0, d2, d4, d6}, [r0]
-void vst4q_f16_ptr(__transfersize(32) __fp16 * ptr, float16x8x4_t * val);
+_NEON2SSESTORAGE void vst4q_f16_ptr(__transfersize(32) __fp16 * ptr, float16x8x4_t * val);
 // IA32 SIMD doesn't work with 16bit floats currently
 
 //void vst4q_f32(__transfersize(16) float32_t * ptr, float32x4x4_t val)// VST4.32 {d0, d2, d4, d6}, [r0]
@@ -11519,14 +11521,14 @@ _NEON2SSE_INLINE void vst4q_f32_ptr(__transfersize(16) float32_t * ptr, float32x
 #define vst4q_f32(ptr, val) vst4q_f32_ptr(ptr, &val)
 
 //void vst4q_p8(__transfersize(64) poly8_t * ptr, poly8x16x4_t val);// VST4.8 {d0, d2, d4, d6}, [r0]
-void vst4q_p8_ptr(__transfersize(64) poly8_t * ptr, poly8x16x4_t * val);
+_NEON2SSESTORAGE void vst4q_p8_ptr(__transfersize(64) poly8_t * ptr, poly8x16x4_t * val);
 #define vst4q_p8 vst4q_u8
 
 //void vst4q_p16(__transfersize(32) poly16_t * ptr, poly16x8x4_t val);// VST4.16 {d0, d2, d4, d6}, [r0]
-void vst4q_p16_ptr(__transfersize(32) poly16_t * ptr, poly16x8x4_t * val);
+_NEON2SSESTORAGE void vst4q_p16_ptr(__transfersize(32) poly16_t * ptr, poly16x8x4_t * val);
 #define vst4q_p16 vst4q_s16
 
-void vst4_u8(__transfersize(32) uint8_t * ptr, uint8x8x4_t val);// VST4.8 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_u8(__transfersize(32) uint8_t * ptr, uint8x8x4_t val);// VST4.8 {d0, d1, d2, d3}, [r0]
 _NEON2SSE_INLINE void vst4_u8(__transfersize(32) uint8_t * ptr, uint8x8x4_t val)
 {
     __m128i sh0, sh1, val0, val2;
@@ -11538,7 +11540,7 @@ _NEON2SSE_INLINE void vst4_u8(__transfersize(32) uint8_t * ptr, uint8x8x4_t val)
     vst1q_u8((ptr + 16),  val2);
 }
 
-void vst4_u16(__transfersize(16) uint16_t * ptr, uint16x4x4_t val);// VST4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_u16(__transfersize(16) uint16_t * ptr, uint16x4x4_t val);// VST4.16 {d0, d1, d2, d3}, [r0]
 _NEON2SSE_INLINE void vst4_u16(__transfersize(16) uint16_t * ptr, uint16x4x4_t val)
 {
     __m128i sh0, sh1, val0, val2;
@@ -11550,7 +11552,7 @@ _NEON2SSE_INLINE void vst4_u16(__transfersize(16) uint16_t * ptr, uint16x4x4_t v
     vst1q_u16((ptr + 8),  val2);
 }
 
-void vst4_u32(__transfersize(8) uint32_t * ptr, uint32x2x4_t val);// VST4.32 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_u32(__transfersize(8) uint32_t * ptr, uint32x2x4_t val);// VST4.32 {d0, d1, d2, d3}, [r0]
 _NEON2SSE_INLINE void vst4_u32(__transfersize(8) uint32_t * ptr, uint32x2x4_t val)
 {
     //0,4,   1,5,  2,6,  3,7
@@ -11563,7 +11565,7 @@ _NEON2SSE_INLINE void vst4_u32(__transfersize(8) uint32_t * ptr, uint32x2x4_t va
     vst1q_u32((ptr + 4),  val1);
 }
 
-void vst4_u64(__transfersize(4) uint64_t * ptr, uint64x1x4_t val);// VST1.64 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_u64(__transfersize(4) uint64_t * ptr, uint64x1x4_t val);// VST1.64 {d0, d1, d2, d3}, [r0]
 _NEON2SSE_INLINE void vst4_u64(__transfersize(4) uint64_t * ptr, uint64x1x4_t val)
 {
     *(ptr) =  val.val[0].m64_u64[0];
@@ -11582,14 +11584,14 @@ _NEON2SSE_INLINE void vst4_u64(__transfersize(4) uint64_t * ptr, uint64x1x4_t va
 #define vst4_s32(ptr, val) vst4_u32((uint32_t*)ptr, val)
 
 //void vst4_s64(__transfersize(4) int64_t * ptr, int64x1x4_t val); // VST1.64 {d0, d1, d2, d3}, [r0]
-void vst4_s64_ptr(__transfersize(4) int64_t * ptr, int64x1x4_t * val);
+_NEON2SSESTORAGE void vst4_s64_ptr(__transfersize(4) int64_t * ptr, int64x1x4_t * val);
 #define vst4_s64(ptr, val) vst4_u64((uint64_t*)ptr, val)
 
 //void vst4_f16(__transfersize(16) __fp16 * ptr, float16x4x4_t val);// VST4.16 {d0, d1, d2, d3}, [r0]
-void vst4_f16_ptr(__transfersize(16) __fp16 * ptr, float16x4x4_t * val);
+_NEON2SSESTORAGE void vst4_f16_ptr(__transfersize(16) __fp16 * ptr, float16x4x4_t * val);
 // IA32 SIMD doesn't work with 16bit floats currently, so need to go to 32 bit and then work with two 128bit registers. See vld1q_f16 for example
 
-void vst4_f32(__transfersize(8) float32_t * ptr, float32x2x4_t val);// VST4.32 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_f32(__transfersize(8) float32_t * ptr, float32x2x4_t val);// VST4.32 {d0, d1, d2, d3}, [r0]
 _NEON2SSE_INLINE void vst4_f32(__transfersize(8) float32_t * ptr, float32x2x4_t val)
 {
     //0,4,   1,5,  2,6,  3,7 -> 0,1, 2,3, 4,5, 6,7
@@ -11603,10 +11605,10 @@ _NEON2SSE_INLINE void vst4_f32(__transfersize(8) float32_t * ptr, float32x2x4_t
     *(ptr + 7) = val.val[3].m64_f32[1];
 }
 
-void vst4_p8(__transfersize(32) poly8_t * ptr, poly8x8x4_t val);// VST4.8 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_p8(__transfersize(32) poly8_t * ptr, poly8x8x4_t val);// VST4.8 {d0, d1, d2, d3}, [r0]
 #define vst4_p8 vst4_u8
 
-void vst4_p16(__transfersize(16) poly16_t * ptr, poly16x4x4_t val);// VST4.16 {d0, d1, d2, d3}, [r0]
+_NEON2SSESTORAGE void vst4_p16(__transfersize(16) poly16_t * ptr, poly16x4x4_t val);// VST4.16 {d0, d1, d2, d3}, [r0]
 #define vst4_p16 vst4_u16
 
 //*********** Store a lane of a vector into memory (extract given lane) for a couple of vectors  *********************
@@ -11628,15 +11630,15 @@ _NEON2SSE_INLINE void vst2q_lane_u32_ptr(__transfersize(2) uint32_t* ptr, uint32
 #define vst2q_lane_u32(ptr, val, lane) vst2q_lane_u32_ptr(ptr, &val, lane)
 
 //void vst2q_lane_s16(__transfersize(2) int16_t * ptr, int16x8x2_t val, __constrange(0,7) int lane);// VST2.16 {d0[0], d2[0]}, [r0]
-void vst2q_lane_s16_ptr(__transfersize(2) int16_t * ptr, int16x8x2_t * val, __constrange(0,7) int lane);
+_NEON2SSESTORAGE void vst2q_lane_s16_ptr(__transfersize(2) int16_t * ptr, int16x8x2_t * val, __constrange(0,7) int lane);
 #define vst2q_lane_s16(ptr, val, lane) vst2q_lane_u16((uint16_t*)ptr, val, lane)
 
 //void vst2q_lane_s32(__transfersize(2) int32_t * ptr, int32x4x2_t val, __constrange(0,3) int lane);// VST2.32 {d0[0], d2[0]}, [r0]
-void vst2q_lane_s32_ptr(__transfersize(2) int32_t * ptr, int32x4x2_t * val, __constrange(0,3) int lane);
+_NEON2SSESTORAGE void vst2q_lane_s32_ptr(__transfersize(2) int32_t * ptr, int32x4x2_t * val, __constrange(0,3) int lane);
 #define vst2q_lane_s32(ptr, val, lane)  vst2q_lane_u32((uint32_t*)ptr, val, lane)
 
 //void vst2q_lane_f16(__transfersize(2) __fp16 * ptr, float16x8x2_t val, __constrange(0,7) int lane);// VST2.16 {d0[0], d2[0]}, [r0]
-void vst2q_lane_f16_ptr(__transfersize(2) __fp16 * ptr, float16x8x2_t * val, __constrange(0,7) int lane);
+_NEON2SSESTORAGE void vst2q_lane_f16_ptr(__transfersize(2) __fp16 * ptr, float16x8x2_t * val, __constrange(0,7) int lane);
 //current IA SIMD doesn't support float16
 
 //void vst2q_lane_f32(__transfersize(2) float32_t * ptr, float32x4x2_t val, __constrange(0,3) int lane)// VST2.32 {d0[0], d2[0]}, [r0]
@@ -11648,53 +11650,53 @@ _NEON2SSE_INLINE void vst2q_lane_f32_ptr(__transfersize(2) float32_t* ptr, float
 #define vst2q_lane_f32(ptr,src,lane) vst2q_lane_f32_ptr(ptr,&src,lane)
 
 //void vst2q_lane_p16(__transfersize(2) poly16_t * ptr, poly16x8x2_t val, __constrange(0,7) int lane);// VST2.16 {d0[0], d2[0]}, [r0]
-void vst2q_lane_p16_ptr(__transfersize(2) poly16_t * ptr, poly16x8x2_t * val, __constrange(0,7) int lane);
+_NEON2SSESTORAGE void vst2q_lane_p16_ptr(__transfersize(2) poly16_t * ptr, poly16x8x2_t * val, __constrange(0,7) int lane);
 #define vst2q_lane_p16 vst2q_lane_s16
 
-void vst2_lane_u8(__transfersize(2) uint8_t * ptr, uint8x8x2_t val, __constrange(0,7) int lane);// VST2.8 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_u8(__transfersize(2) uint8_t * ptr, uint8x8x2_t val, __constrange(0,7) int lane);// VST2.8 {d0[0], d1[0]}, [r0]
 _NEON2SSE_INLINE void vst2_lane_u8(__transfersize(2) uint8_t * ptr, uint8x8x2_t val, __constrange(0,7) int lane) // VST2.8 {d0[0], d1[0]}, [r0]
 {
     *(ptr) = val.val[0].m64_u8[lane];
     *(ptr + 1) = val.val[1].m64_u8[lane];
 }
 
-void vst2_lane_u16(__transfersize(2) uint16_t * ptr, uint16x4x2_t val, __constrange(0,3) int lane);// VST2.16 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_u16(__transfersize(2) uint16_t * ptr, uint16x4x2_t val, __constrange(0,3) int lane);// VST2.16 {d0[0], d1[0]}, [r0]
 _NEON2SSE_INLINE void vst2_lane_u16(__transfersize(2) uint16_t * ptr, uint16x4x2_t val, __constrange(0,3) int lane)
 {
     *(ptr) = val.val[0].m64_u16[lane];
     *(ptr + 1) = val.val[1].m64_u16[lane];
 }
 
-void vst2_lane_u32(__transfersize(2) uint32_t * ptr, uint32x2x2_t val, __constrange(0,1) int lane);// VST2.32 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_u32(__transfersize(2) uint32_t * ptr, uint32x2x2_t val, __constrange(0,1) int lane);// VST2.32 {d0[0], d1[0]}, [r0]
 _NEON2SSE_INLINE void vst2_lane_u32(__transfersize(2) uint32_t * ptr, uint32x2x2_t val, __constrange(0,1) int lane)
 {
     *(ptr) = val.val[0].m64_u32[lane];
     *(ptr + 1) = val.val[1].m64_u32[lane];
 }
 
-void vst2_lane_s8(__transfersize(2) int8_t * ptr, int8x8x2_t val, __constrange(0,7) int lane);// VST2.8 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_s8(__transfersize(2) int8_t * ptr, int8x8x2_t val, __constrange(0,7) int lane);// VST2.8 {d0[0], d1[0]}, [r0]
 #define vst2_lane_s8(ptr, val, lane)  vst2_lane_u8((uint8_t*)ptr, val, lane)
 
-void vst2_lane_s16(__transfersize(2) int16_t * ptr, int16x4x2_t val, __constrange(0,3) int lane);// VST2.16 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_s16(__transfersize(2) int16_t * ptr, int16x4x2_t val, __constrange(0,3) int lane);// VST2.16 {d0[0], d1[0]}, [r0]
 #define vst2_lane_s16(ptr, val, lane)  vst2_lane_u16((uint16_t*)ptr, val, lane)
 
-void vst2_lane_s32(__transfersize(2) int32_t * ptr, int32x2x2_t val, __constrange(0,1) int lane);// VST2.32 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_s32(__transfersize(2) int32_t * ptr, int32x2x2_t val, __constrange(0,1) int lane);// VST2.32 {d0[0], d1[0]}, [r0]
 #define vst2_lane_s32(ptr, val, lane)  vst2_lane_u32((uint32_t*)ptr, val, lane)
 
 //void vst2_lane_f16(__transfersize(2) __fp16 * ptr, float16x4x2_t val, __constrange(0,3) int lane); // VST2.16 {d0[0], d1[0]}, [r0]
 //current IA SIMD doesn't support float16
 
-void vst2_lane_f32(__transfersize(2) float32_t * ptr, float32x2x2_t val, __constrange(0,1) int lane); // VST2.32 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_f32(__transfersize(2) float32_t * ptr, float32x2x2_t val, __constrange(0,1) int lane); // VST2.32 {d0[0], d1[0]}, [r0]
 _NEON2SSE_INLINE void vst2_lane_f32(__transfersize(2) float32_t * ptr, float32x2x2_t val, __constrange(0,1) int lane)
 {
     *(ptr) = val.val[0].m64_f32[lane];
     *(ptr + 1) = val.val[1].m64_f32[lane];
 }
 
-void vst2_lane_p8(__transfersize(2) poly8_t * ptr, poly8x8x2_t val, __constrange(0,7) int lane);// VST2.8 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_p8(__transfersize(2) poly8_t * ptr, poly8x8x2_t val, __constrange(0,7) int lane);// VST2.8 {d0[0], d1[0]}, [r0]
 #define vst2_lane_p8 vst2_lane_u8
 
-void vst2_lane_p16(__transfersize(2) poly16_t * ptr, poly16x4x2_t val, __constrange(0,3) int lane);// VST2.16 {d0[0], d1[0]}, [r0]
+_NEON2SSESTORAGE void vst2_lane_p16(__transfersize(2) poly16_t * ptr, poly16x4x2_t val, __constrange(0,3) int lane);// VST2.16 {d0[0], d1[0]}, [r0]
 #define vst2_lane_p16 vst2_lane_u16
 
 //************************* Triple lanes  stores *******************************************************
@@ -11716,15 +11718,15 @@ _NEON2SSE_INLINE void vst3q_lane_u32_ptr(__transfersize(3) uint32_t * ptr, uint3
 #define vst3q_lane_u32(ptr, val, lane) vst3q_lane_u32_ptr(ptr, &val, lane)
 
 //void vst3q_lane_s16(__transfersize(3) int16_t * ptr, int16x8x3_t val, __constrange(0,7) int lane);// VST3.16 {d0[0], d2[0], d4[0]}, [r0]
-void vst3q_lane_s16_ptr(__transfersize(3) int16_t * ptr, int16x8x3_t * val, __constrange(0,7) int lane);
+_NEON2SSESTORAGE void vst3q_lane_s16_ptr(__transfersize(3) int16_t * ptr, int16x8x3_t * val, __constrange(0,7) int lane);
 #define vst3q_lane_s16(ptr, val, lane) vst3q_lane_u16((uint16_t *)ptr, val, lane)
 
 //void vst3q_lane_s32(__transfersize(3) int32_t * ptr, int32x4x3_t val, __constrange(0,3) int lane);// VST3.32 {d0[0], d2[0], d4[0]}, [r0]
-void vst3q_lane_s32_ptr(__transfersize(3) int32_t * ptr, int32x4x3_t * val, __constrange(0,3) int lane);
+_NEON2SSESTORAGE void vst3q_lane_s32_ptr(__transfersize(3) int32_t * ptr, int32x4x3_t * val, __constrange(0,3) int lane);
 #define vst3q_lane_s32(ptr, val, lane) vst3q_lane_u32((uint32_t *)ptr, val, lane)
 
 //void vst3q_lane_f16(__transfersize(3) __fp16 * ptr, float16x8x3_t val, __constrange(0,7) int lane);// VST3.16 {d0[0], d2[0], d4[0]}, [r0]
-void vst3q_lane_f16_ptr(__transfersize(3) __fp16 * ptr, float16x8x3_t * val, __constrange(0,7) int lane);
+_NEON2SSESTORAGE void vst3q_lane_f16_ptr(__transfersize(3) __fp16 * ptr, float16x8x3_t * val, __constrange(0,7) int lane);
 //current IA SIMD doesn't support float16
 
 //void vst3q_lane_f32(__transfersize(3) float32_t * ptr, float32x4x3_t val, __constrange(0,3) int lane)// VST3.32 {d0[0], d2[0], d4[0]}, [r0]
@@ -11737,10 +11739,10 @@ _NEON2SSE_INLINE void vst3q_lane_f32_ptr(__transfersize(3) float32_t * ptr, floa
 #define vst3q_lane_f32(ptr,val,lane) vst3q_lane_f32_ptr(ptr,&val,lane)
 
 //void vst3q_lane_p16(__transfersize(3) poly16_t * ptr, poly16x8x3_t val, __constrange(0,7) int lane);// VST3.16 {d0[0], d2[0], d4[0]}, [r0]
-void vst3q_lane_p16_ptr(__transfersize(3) poly16_t * ptr, poly16x8x3_t * val, __constrange(0,7) int lane);
+_NEON2SSESTORAGE void vst3q_lane_p16_ptr(__transfersize(3) poly16_t * ptr, poly16x8x3_t * val, __constrange(0,7) int lane);
 #define vst3q_lane_p16 vst3q_lane_s16
 
-void vst3_lane_u8(__transfersize(3) uint8_t * ptr, uint8x8x3_t val, __constrange(0,7) int lane);// VST3.8 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_u8(__transfersize(3) uint8_t * ptr, uint8x8x3_t val, __constrange(0,7) int lane);// VST3.8 {d0[0], d1[0], d2[0]}, [r0]
 _NEON2SSE_INLINE void vst3_lane_u8(__transfersize(3) uint8_t * ptr, uint8x8x3_t val, __constrange(0,7) int lane)
 {
     *(ptr) =     val.val[0].m64_u8[lane];
@@ -11748,7 +11750,7 @@ _NEON2SSE_INLINE void vst3_lane_u8(__transfersize(3) uint8_t * ptr, uint8x8x3_t
     *(ptr + 2) = val.val[2].m64_u8[lane];
 }
 
-void vst3_lane_u16(__transfersize(3) uint16_t * ptr, uint16x4x3_t val, __constrange(0,3) int lane);// VST3.16 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_u16(__transfersize(3) uint16_t * ptr, uint16x4x3_t val, __constrange(0,3) int lane);// VST3.16 {d0[0], d1[0], d2[0]}, [r0]
 _NEON2SSE_INLINE void vst3_lane_u16(__transfersize(3) uint16_t * ptr, uint16x4x3_t val, __constrange(0,3) int lane)
 {
     *(ptr) =     val.val[0].m64_u16[lane];
@@ -11756,7 +11758,7 @@ _NEON2SSE_INLINE void vst3_lane_u16(__transfersize(3) uint16_t * ptr, uint16x4x3
     *(ptr + 2) = val.val[2].m64_u16[lane];
 }
 
-void vst3_lane_u32(__transfersize(3) uint32_t * ptr, uint32x2x3_t val, __constrange(0,1) int lane);// VST3.32 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_u32(__transfersize(3) uint32_t * ptr, uint32x2x3_t val, __constrange(0,1) int lane);// VST3.32 {d0[0], d1[0], d2[0]}, [r0]
 _NEON2SSE_INLINE void vst3_lane_u32(__transfersize(3) uint32_t * ptr, uint32x2x3_t val, __constrange(0,1) int lane)
 {
     *(ptr) =     val.val[0].m64_u32[lane];
@@ -11764,20 +11766,20 @@ _NEON2SSE_INLINE void vst3_lane_u32(__transfersize(3) uint32_t * ptr, uint32x2x3
     *(ptr + 2) = val.val[2].m64_u32[lane];
 }
 
-void vst3_lane_s8(__transfersize(3) int8_t * ptr, int8x8x3_t val, __constrange(0,7) int lane);// VST3.8 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_s8(__transfersize(3) int8_t * ptr, int8x8x3_t val, __constrange(0,7) int lane);// VST3.8 {d0[0], d1[0], d2[0]}, [r0]
 #define  vst3_lane_s8(ptr, val, lane) vst3_lane_u8((uint8_t *)ptr, val, lane)
 
-void vst3_lane_s16(__transfersize(3) int16_t * ptr, int16x4x3_t val, __constrange(0,3) int lane);// VST3.16 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_s16(__transfersize(3) int16_t * ptr, int16x4x3_t val, __constrange(0,3) int lane);// VST3.16 {d0[0], d1[0], d2[0]}, [r0]
 #define vst3_lane_s16(ptr, val, lane) vst3_lane_u16((uint16_t *)ptr, val, lane)
 
-void vst3_lane_s32(__transfersize(3) int32_t * ptr, int32x2x3_t val, __constrange(0,1) int lane);// VST3.32 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_s32(__transfersize(3) int32_t * ptr, int32x2x3_t val, __constrange(0,1) int lane);// VST3.32 {d0[0], d1[0], d2[0]}, [r0]
 #define vst3_lane_s32(ptr, val, lane) vst3_lane_u32((uint32_t *)ptr, val, lane)
 
 //void vst3_lane_f16(__transfersize(3) __fp16 * ptr, float16x4x3_t val, __constrange(0,3) int lane);// VST3.16 {d0[0], d1[0], d2[0]}, [r0]
-void vst3_lane_f16_ptr(__transfersize(3) __fp16 * ptr, float16x4x3_t * val, __constrange(0,3) int lane);
+_NEON2SSESTORAGE void vst3_lane_f16_ptr(__transfersize(3) __fp16 * ptr, float16x4x3_t * val, __constrange(0,3) int lane);
 //current IA SIMD doesn't support float16
 
-void vst3_lane_f32(__transfersize(3) float32_t * ptr, float32x2x3_t val, __constrange(0,1) int lane);// VST3.32 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_f32(__transfersize(3) float32_t * ptr, float32x2x3_t val, __constrange(0,1) int lane);// VST3.32 {d0[0], d1[0], d2[0]}, [r0]
 _NEON2SSE_INLINE void vst3_lane_f32(__transfersize(3) float32_t * ptr, float32x2x3_t val, __constrange(0,1) int lane)
 {
     *(ptr) = val.val[0].m64_f32[lane];
@@ -11785,10 +11787,10 @@ _NEON2SSE_INLINE void vst3_lane_f32(__transfersize(3) float32_t * ptr, float32x2
     *(ptr + 2) = val.val[2].m64_f32[lane];
 }
 
-void vst3_lane_p8(__transfersize(3) poly8_t * ptr, poly8x8x3_t val, __constrange(0,7) int lane);// VST3.8 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_p8(__transfersize(3) poly8_t * ptr, poly8x8x3_t val, __constrange(0,7) int lane);// VST3.8 {d0[0], d1[0], d2[0]}, [r0]
 #define vst3_lane_p8 vst3_lane_u8
 
-void vst3_lane_p16(__transfersize(3) poly16_t * ptr, poly16x4x3_t val, __constrange(0,3) int lane);// VST3.16 {d0[0], d1[0], d2[0]}, [r0]
+_NEON2SSESTORAGE void vst3_lane_p16(__transfersize(3) poly16_t * ptr, poly16x4x3_t val, __constrange(0,3) int lane);// VST3.16 {d0[0], d1[0], d2[0]}, [r0]
 #define vst3_lane_p16 vst3_lane_u16
 
 //******************************** Quadruple lanes stores ***********************************************
@@ -11810,15 +11812,15 @@ _NEON2SSE_INLINE void vst4q_lane_u32_ptr(__transfersize(4) uint32_t * ptr, uint3
 #define vst4q_lane_u32(ptr, val, lane) vst4q_lane_u32_ptr(ptr, &val, lane)
 
 //void vst4q_lane_s16(__transfersize(4) int16_t * ptr, int16x8x4_t val, __constrange(0,7) int lane);// VST4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-void vst4q_lane_s16_ptr(__transfersize(4) int16_t * ptr, int16x8x4_t * val, __constrange(0,7) int lane);
+_NEON2SSESTORAGE void vst4q_lane_s16_ptr(__transfersize(4) int16_t * ptr, int16x8x4_t * val, __constrange(0,7) int lane);
 #define vst4q_lane_s16(ptr,val,lane) vst4q_lane_u16((uint16_t *)ptr,val,lane)
 
 //void vst4q_lane_s32(__transfersize(4) int32_t * ptr, int32x4x4_t val, __constrange(0,3) int lane);// VST4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-void vst4q_lane_s32_ptr(__transfersize(4) int32_t * ptr, int32x4x4_t * val, __constrange(0,3) int lane);
+_NEON2SSESTORAGE void vst4q_lane_s32_ptr(__transfersize(4) int32_t * ptr, int32x4x4_t * val, __constrange(0,3) int lane);
 #define vst4q_lane_s32(ptr,val,lane) vst4q_lane_u32((uint32_t *)ptr,val,lane)
 
 //void vst4q_lane_f16(__transfersize(4) __fp16 * ptr, float16x8x4_t val, __constrange(0,7) int lane);// VST4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-void vst4q_lane_f16_ptr(__transfersize(4) __fp16 * ptr, float16x8x4_t * val, __constrange(0,7) int lane);
+_NEON2SSESTORAGE void vst4q_lane_f16_ptr(__transfersize(4) __fp16 * ptr, float16x8x4_t * val, __constrange(0,7) int lane);
 //current IA SIMD doesn't support float16
 
 //void vst4q_lane_f32(__transfersize(4) float32_t * ptr, float32x4x4_t val, __constrange(0,3) int lane)// VST4.32 {d0[0], d2[0], d4[0], d6[0]}, [r0]
@@ -11832,10 +11834,10 @@ _NEON2SSE_INLINE void vst4q_lane_f32_ptr(__transfersize(4) float32_t * ptr, floa
 #define vst4q_lane_f32(ptr,val,lane) vst4q_lane_f32_ptr(ptr,&val,lane)
 
 //void vst4q_lane_p16(__transfersize(4) poly16_t * ptr, poly16x8x4_t val, __constrange(0,7) int lane);// VST4.16 {d0[0], d2[0], d4[0], d6[0]}, [r0]
-void vst4q_lane_p16_ptr(__transfersize(4) poly16_t * ptr, poly16x8x4_t * val, __constrange(0,7) int lane);
+_NEON2SSESTORAGE void vst4q_lane_p16_ptr(__transfersize(4) poly16_t * ptr, poly16x8x4_t * val, __constrange(0,7) int lane);
 #define vst4q_lane_p16 vst4q_lane_u16
 
-void vst4_lane_u8(__transfersize(4) uint8_t * ptr, uint8x8x4_t val, __constrange(0,7) int lane);// VST4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_u8(__transfersize(4) uint8_t * ptr, uint8x8x4_t val, __constrange(0,7) int lane);// VST4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 _NEON2SSE_INLINE void vst4_lane_u8(__transfersize(4) uint8_t * ptr, uint8x8x4_t val, __constrange(0,7) int lane)
 {
     *(ptr) =     val.val[0].m64_u8[lane];
@@ -11844,7 +11846,7 @@ _NEON2SSE_INLINE void vst4_lane_u8(__transfersize(4) uint8_t * ptr, uint8x8x4_t
     *(ptr + 3) = val.val[3].m64_u8[lane];
 }
 
-void vst4_lane_u16(__transfersize(4) uint16_t * ptr, uint16x4x4_t val, __constrange(0,3) int lane);// VST4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_u16(__transfersize(4) uint16_t * ptr, uint16x4x4_t val, __constrange(0,3) int lane);// VST4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 _NEON2SSE_INLINE void vst4_lane_u16(__transfersize(4) uint16_t * ptr, uint16x4x4_t val, __constrange(0,3) int lane)
 {
     *(ptr) =     val.val[0].m64_u16[lane];
@@ -11853,7 +11855,7 @@ _NEON2SSE_INLINE void vst4_lane_u16(__transfersize(4) uint16_t * ptr, uint16x4x4
     *(ptr + 3) = val.val[3].m64_u16[lane];
 }
 
-void vst4_lane_u32(__transfersize(4) uint32_t * ptr, uint32x2x4_t val, __constrange(0,1) int lane);// VST4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_u32(__transfersize(4) uint32_t * ptr, uint32x2x4_t val, __constrange(0,1) int lane);// VST4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 _NEON2SSE_INLINE void vst4_lane_u32(__transfersize(4) uint32_t * ptr, uint32x2x4_t val, __constrange(0,1) int lane)
 {
     *(ptr) =     val.val[0].m64_u32[lane];
@@ -11862,20 +11864,20 @@ _NEON2SSE_INLINE void vst4_lane_u32(__transfersize(4) uint32_t * ptr, uint32x2x4
     *(ptr + 3) = val.val[3].m64_u32[lane];
 }
 
-void vst4_lane_s8(__transfersize(4) int8_t * ptr, int8x8x4_t val, __constrange(0,7) int lane);// VST4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_s8(__transfersize(4) int8_t * ptr, int8x8x4_t val, __constrange(0,7) int lane);// VST4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 #define vst4_lane_s8(ptr, val, lane) vst4_lane_u8((uint8_t*)ptr, val, lane)
 
-void vst4_lane_s16(__transfersize(4) int16_t * ptr, int16x4x4_t val, __constrange(0,3) int lane);// VST4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_s16(__transfersize(4) int16_t * ptr, int16x4x4_t val, __constrange(0,3) int lane);// VST4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 #define vst4_lane_s16(ptr, val, lane) vst4_lane_u16((uint16_t*)ptr, val, lane)
 
-void vst4_lane_s32(__transfersize(4) int32_t * ptr, int32x2x4_t val, __constrange(0,1) int lane);// VST4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_s32(__transfersize(4) int32_t * ptr, int32x2x4_t val, __constrange(0,1) int lane);// VST4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 #define vst4_lane_s32(ptr, val, lane) vst4_lane_u32((uint32_t*)ptr, val, lane)
 
 //void vst4_lane_f16(__transfersize(4) __fp16 * ptr, float16x4x4_t val, __constrange(0,3) int lane);// VST4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
-void vst4_lane_f16_ptr(__transfersize(4) __fp16 * ptr, float16x4x4_t * val, __constrange(0,3) int lane);
+_NEON2SSESTORAGE void vst4_lane_f16_ptr(__transfersize(4) __fp16 * ptr, float16x4x4_t * val, __constrange(0,3) int lane);
 //current IA SIMD doesn't support float16
 
-void vst4_lane_f32(__transfersize(4) float32_t * ptr, float32x2x4_t  val, __constrange(0,1) int lane); // VST4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_f32(__transfersize(4) float32_t * ptr, float32x2x4_t  val, __constrange(0,1) int lane); // VST4.32 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 _NEON2SSE_INLINE void vst4_lane_f32(__transfersize(4) float32_t * ptr, float32x2x4_t val, __constrange(0,1) int lane)
 {
     *(ptr) = val.val[0].m64_f32[lane];
@@ -11884,69 +11886,69 @@ _NEON2SSE_INLINE void vst4_lane_f32(__transfersize(4) float32_t * ptr, float32x2
     *(ptr + 3) = val.val[3].m64_f32[lane];
 }
 
-void vst4_lane_p8(__transfersize(4) poly8_t * ptr, poly8x8x4_t val, __constrange(0,7) int lane);// VST4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_p8(__transfersize(4) poly8_t * ptr, poly8x8x4_t val, __constrange(0,7) int lane);// VST4.8 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 #define vst4_lane_p8 vst4_lane_u8
 
-void vst4_lane_p16(__transfersize(4) poly16_t * ptr, poly16x4x4_t val, __constrange(0,3) int lane);// VST4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
+_NEON2SSESTORAGE void vst4_lane_p16(__transfersize(4) poly16_t * ptr, poly16x4x4_t val, __constrange(0,3) int lane);// VST4.16 {d0[0], d1[0], d2[0], d3[0]}, [r0]
 #define vst4_lane_p16 vst4_lane_u16
 
 //**************************************************************************************************
 //************************ Extract lanes from a vector ********************************************
 //**************************************************************************************************
 //These intrinsics extract a single lane (element) from a vector.
-uint8_t vget_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VMOV.U8 r0, d0[0]
+_NEON2SSESTORAGE uint8_t vget_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VMOV.U8 r0, d0[0]
 #define vget_lane_u8(vec, lane) vec.m64_u8[lane]
 
-uint16_t vget_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VMOV.s16 r0, d0[0]
+_NEON2SSESTORAGE uint16_t vget_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VMOV.s16 r0, d0[0]
 #define vget_lane_u16(vec, lane) vec.m64_u16[lane]
 
 
-uint32_t vget_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE uint32_t vget_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
 #define vget_lane_u32(vec, lane) vec.m64_u32[lane]
 
-int8_t vget_lane_s8(int8x8_t vec, __constrange(0,7) int lane); // VMOV.S8 r0, d0[0]
+_NEON2SSESTORAGE int8_t vget_lane_s8(int8x8_t vec, __constrange(0,7) int lane); // VMOV.S8 r0, d0[0]
 #define vget_lane_s8(vec, lane) vec.m64_i8[lane]
 
-int16_t vget_lane_s16(int16x4_t vec, __constrange(0,3) int lane); // VMOV.S16 r0, d0[0]
+_NEON2SSESTORAGE int16_t vget_lane_s16(int16x4_t vec, __constrange(0,3) int lane); // VMOV.S16 r0, d0[0]
 #define vget_lane_s16(vec, lane) vec.m64_i16[lane]
 
-int32_t vget_lane_s32(int32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE int32_t vget_lane_s32(int32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
 #define vget_lane_s32(vec, lane) vec.m64_i32[lane]
 
-poly8_t vget_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VMOV.U8 r0, d0[0]
+_NEON2SSESTORAGE poly8_t vget_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VMOV.U8 r0, d0[0]
 #define vget_lane_p8 vget_lane_u8
 
-poly16_t vget_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VMOV.s16 r0, d0[0]
+_NEON2SSESTORAGE poly16_t vget_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VMOV.s16 r0, d0[0]
 #define vget_lane_p16 vget_lane_u16
 
-float32_t vget_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE float32_t vget_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VMOV.32 r0, d0[0]
 #define vget_lane_f32(vec, lane) vec.m64_f32[lane]
 
-uint8_t vgetq_lane_u8(uint8x16_t vec, __constrange(0,15) int lane); // VMOV.U8 r0, d0[0]
+_NEON2SSESTORAGE uint8_t vgetq_lane_u8(uint8x16_t vec, __constrange(0,15) int lane); // VMOV.U8 r0, d0[0]
 #define vgetq_lane_u8 (uint8_t) _MM_EXTRACT_EPI8
 
-uint16_t vgetq_lane_u16(uint16x8_t vec, __constrange(0,7) int lane); // VMOV.s16 r0, d0[0]
+_NEON2SSESTORAGE uint16_t vgetq_lane_u16(uint16x8_t vec, __constrange(0,7) int lane); // VMOV.s16 r0, d0[0]
 #define  vgetq_lane_u16 (uint16_t) _MM_EXTRACT_EPI16
 
-uint32_t vgetq_lane_u32(uint32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE uint32_t vgetq_lane_u32(uint32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
 #define vgetq_lane_u32 (uint32_t) _MM_EXTRACT_EPI32
 
-int8_t vgetq_lane_s8(int8x16_t vec, __constrange(0,15) int lane); // VMOV.S8 r0, d0[0]
+_NEON2SSESTORAGE int8_t vgetq_lane_s8(int8x16_t vec, __constrange(0,15) int lane); // VMOV.S8 r0, d0[0]
 #define vgetq_lane_s8 _MM_EXTRACT_EPI8
 
-int16_t vgetq_lane_s16(int16x8_t vec, __constrange(0,7) int lane); // VMOV.S16 r0, d0[0]
+_NEON2SSESTORAGE int16_t vgetq_lane_s16(int16x8_t vec, __constrange(0,7) int lane); // VMOV.S16 r0, d0[0]
 #define vgetq_lane_s16 _MM_EXTRACT_EPI16
 
-int32_t vgetq_lane_s32(int32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE int32_t vgetq_lane_s32(int32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
 #define vgetq_lane_s32 _MM_EXTRACT_EPI32
 
-poly8_t vgetq_lane_p8(poly8x16_t vec, __constrange(0,15) int lane); // VMOV.U8 r0, d0[0]
+_NEON2SSESTORAGE poly8_t vgetq_lane_p8(poly8x16_t vec, __constrange(0,15) int lane); // VMOV.U8 r0, d0[0]
 #define vgetq_lane_p8 vgetq_lane_u8
 
-poly16_t vgetq_lane_p16(poly16x8_t vec, __constrange(0,7) int lane); // VMOV.s16 r0, d0[0]
+_NEON2SSESTORAGE poly16_t vgetq_lane_p16(poly16x8_t vec, __constrange(0,7) int lane); // VMOV.s16 r0, d0[0]
 #define vgetq_lane_p16 vgetq_lane_u16
 
-float32_t vgetq_lane_f32(float32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
+_NEON2SSESTORAGE float32_t vgetq_lane_f32(float32x4_t vec, __constrange(0,3) int lane); // VMOV.32 r0, d0[0]
 _NEON2SSE_INLINE float32_t vgetq_lane_f32(float32x4_t vec, __constrange(0,3) int lane)
 {
     int32_t ilane;
@@ -11954,17 +11956,17 @@ _NEON2SSE_INLINE float32_t vgetq_lane_f32(float32x4_t vec, __constrange(0,3) int
     return *(float*)&ilane;
 }
 
-int64_t vget_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV r0,r0,d0
+_NEON2SSESTORAGE int64_t vget_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV r0,r0,d0
 #define vget_lane_s64(vec, lane) vec.m64_i64[0]
 
-uint64_t vget_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV r0,r0,d0
+_NEON2SSESTORAGE uint64_t vget_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV r0,r0,d0
 #define vget_lane_u64(vec, lane) vec.m64_u64[0]
 
 
-int64_t vgetq_lane_s64(int64x2_t vec, __constrange(0,1) int lane); // VMOV r0,r0,d0
+_NEON2SSESTORAGE int64_t vgetq_lane_s64(int64x2_t vec, __constrange(0,1) int lane); // VMOV r0,r0,d0
 #define vgetq_lane_s64 _MM_EXTRACT_EPI64
 
-uint64_t vgetq_lane_u64(uint64x2_t vec, __constrange(0,1) int lane); // VMOV r0,r0,d0
+_NEON2SSESTORAGE uint64_t vgetq_lane_u64(uint64x2_t vec, __constrange(0,1) int lane); // VMOV r0,r0,d0
 #define vgetq_lane_u64 (uint64_t) _MM_EXTRACT_EPI64
 
 // ***************** Set lanes within a vector ********************************************
@@ -11972,7 +11974,7 @@ uint64_t vgetq_lane_u64(uint64x2_t vec, __constrange(0,1) int lane); // VMOV r0,
 //These intrinsics set a single lane (element) within a vector.
 //same functions as vld1_lane_xx ones, but take the value to be set directly.
 
-uint8x8_t vset_lane_u8(uint8_t value, uint8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE uint8x8_t vset_lane_u8(uint8_t value, uint8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
 _NEON2SSE_INLINE uint8x8_t vset_lane_u8(uint8_t value, uint8x8_t vec, __constrange(0,7) int lane)
 {
     uint8_t val;
@@ -11980,7 +11982,7 @@ _NEON2SSE_INLINE uint8x8_t vset_lane_u8(uint8_t value, uint8x8_t vec, __constran
     return vld1_lane_u8(&val, vec,  lane);
 }
 
-uint16x4_t vset_lane_u16(uint16_t value, uint16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE uint16x4_t vset_lane_u16(uint16_t value, uint16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
 _NEON2SSE_INLINE uint16x4_t vset_lane_u16(uint16_t value, uint16x4_t vec, __constrange(0,3) int lane)
 {
     uint16_t val;
@@ -11988,7 +11990,7 @@ _NEON2SSE_INLINE uint16x4_t vset_lane_u16(uint16_t value, uint16x4_t vec, __cons
     return vld1_lane_u16(&val, vec,  lane);
 }
 
-uint32x2_t vset_lane_u32(uint32_t value, uint32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE uint32x2_t vset_lane_u32(uint32_t value, uint32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
 _NEON2SSE_INLINE uint32x2_t vset_lane_u32(uint32_t value, uint32x2_t vec, __constrange(0,1) int lane)
 {
     uint32_t val;
@@ -11996,7 +11998,7 @@ _NEON2SSE_INLINE uint32x2_t vset_lane_u32(uint32_t value, uint32x2_t vec, __cons
     return vld1_lane_u32(&val, vec,  lane);
 }
 
-int8x8_t vset_lane_s8(int8_t value, int8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE int8x8_t vset_lane_s8(int8_t value, int8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
 _NEON2SSE_INLINE int8x8_t vset_lane_s8(int8_t value, int8x8_t vec, __constrange(0,7) int lane)
 {
     int8_t val;
@@ -12004,7 +12006,7 @@ _NEON2SSE_INLINE int8x8_t vset_lane_s8(int8_t value, int8x8_t vec, __constrange(
     return vld1_lane_s8(&val, vec,  lane);
 }
 
-int16x4_t vset_lane_s16(int16_t value, int16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE int16x4_t vset_lane_s16(int16_t value, int16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
 _NEON2SSE_INLINE int16x4_t vset_lane_s16(int16_t value, int16x4_t vec, __constrange(0,3) int lane)
 {
     int16_t val;
@@ -12012,7 +12014,7 @@ _NEON2SSE_INLINE int16x4_t vset_lane_s16(int16_t value, int16x4_t vec, __constra
     return vld1_lane_s16(&val, vec,  lane);
 }
 
-int32x2_t vset_lane_s32(int32_t value, int32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE int32x2_t vset_lane_s32(int32_t value, int32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
 _NEON2SSE_INLINE int32x2_t vset_lane_s32(int32_t value, int32x2_t vec, __constrange(0,1) int lane)
 {
     int32_t val;
@@ -12020,13 +12022,13 @@ _NEON2SSE_INLINE int32x2_t vset_lane_s32(int32_t value, int32x2_t vec, __constra
     return vld1_lane_s32(&val, vec,  lane);
 }
 
-poly8x8_t vset_lane_p8(poly8_t value, poly8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE poly8x8_t vset_lane_p8(poly8_t value, poly8x8_t vec, __constrange(0,7) int lane); // VMOV.8 d0[0],r0
 #define vset_lane_p8  vset_lane_u8
 
-poly16x4_t vset_lane_p16(poly16_t value, poly16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE poly16x4_t vset_lane_p16(poly16_t value, poly16x4_t vec, __constrange(0,3) int lane); // VMOV.16 d0[0],r0
 #define vset_lane_p16  vset_lane_u16
 
-float32x2_t vset_lane_f32(float32_t value, float32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE float32x2_t vset_lane_f32(float32_t value, float32x2_t vec, __constrange(0,1) int lane); // VMOV.32 d0[0],r0
 _NEON2SSE_INLINE float32x2_t vset_lane_f32(float32_t value, float32x2_t vec, __constrange(0,1) int lane)
 {
     float32_t val;
@@ -12034,7 +12036,7 @@ _NEON2SSE_INLINE float32x2_t vset_lane_f32(float32_t value, float32x2_t vec, __c
     return vld1_lane_f32(&val, vec,  lane);
 }
 
-uint8x16_t vsetq_lane_u8(uint8_t value, uint8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE uint8x16_t vsetq_lane_u8(uint8_t value, uint8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
 _NEON2SSE_INLINE uint8x16_t vsetq_lane_u8(uint8_t value, uint8x16_t vec, __constrange(0,15) int lane)
 {
     uint8_t val;
@@ -12042,7 +12044,7 @@ _NEON2SSE_INLINE uint8x16_t vsetq_lane_u8(uint8_t value, uint8x16_t vec, __const
     return vld1q_lane_u8(&val, vec,  lane);
 }
 
-uint16x8_t vsetq_lane_u16(uint16_t value, uint16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE uint16x8_t vsetq_lane_u16(uint16_t value, uint16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
 _NEON2SSE_INLINE uint16x8_t vsetq_lane_u16(uint16_t value, uint16x8_t vec, __constrange(0,7) int lane)
 {
     uint16_t val;
@@ -12050,7 +12052,7 @@ _NEON2SSE_INLINE uint16x8_t vsetq_lane_u16(uint16_t value, uint16x8_t vec, __con
     return vld1q_lane_u16(&val, vec,  lane);
 }
 
-uint32x4_t vsetq_lane_u32(uint32_t value, uint32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE uint32x4_t vsetq_lane_u32(uint32_t value, uint32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
 _NEON2SSE_INLINE uint32x4_t vsetq_lane_u32(uint32_t value, uint32x4_t vec, __constrange(0,3) int lane)
 {
     uint32_t val;
@@ -12058,7 +12060,7 @@ _NEON2SSE_INLINE uint32x4_t vsetq_lane_u32(uint32_t value, uint32x4_t vec, __con
     return vld1q_lane_u32(&val, vec,  lane);
 }
 
-int8x16_t vsetq_lane_s8(int8_t value, int8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE int8x16_t vsetq_lane_s8(int8_t value, int8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
 _NEON2SSE_INLINE int8x16_t vsetq_lane_s8(int8_t value, int8x16_t vec, __constrange(0,15) int lane)
 {
     int8_t val;
@@ -12066,7 +12068,7 @@ _NEON2SSE_INLINE int8x16_t vsetq_lane_s8(int8_t value, int8x16_t vec, __constran
     return vld1q_lane_s8(&val, vec,  lane);
 }
 
-int16x8_t vsetq_lane_s16(int16_t value, int16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE int16x8_t vsetq_lane_s16(int16_t value, int16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
 _NEON2SSE_INLINE int16x8_t vsetq_lane_s16(int16_t value, int16x8_t vec, __constrange(0,7) int lane)
 {
     int16_t val;
@@ -12074,7 +12076,7 @@ _NEON2SSE_INLINE int16x8_t vsetq_lane_s16(int16_t value, int16x8_t vec, __constr
     return vld1q_lane_s16(&val, vec,  lane);
 }
 
-int32x4_t vsetq_lane_s32(int32_t value, int32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE int32x4_t vsetq_lane_s32(int32_t value, int32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
 _NEON2SSE_INLINE int32x4_t vsetq_lane_s32(int32_t value, int32x4_t vec, __constrange(0,3) int lane)
 {
     int32_t val;
@@ -12082,13 +12084,13 @@ _NEON2SSE_INLINE int32x4_t vsetq_lane_s32(int32_t value, int32x4_t vec, __constr
     return vld1q_lane_s32(&val, vec,  lane);
 }
 
-poly8x16_t vsetq_lane_p8(poly8_t value, poly8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
+_NEON2SSESTORAGE poly8x16_t vsetq_lane_p8(poly8_t value, poly8x16_t vec, __constrange(0,15) int lane); // VMOV.8 d0[0],r0
 #define vsetq_lane_p8 vsetq_lane_u8
 
-poly16x8_t vsetq_lane_p16(poly16_t value, poly16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
+_NEON2SSESTORAGE poly16x8_t vsetq_lane_p16(poly16_t value, poly16x8_t vec, __constrange(0,7) int lane); // VMOV.16 d0[0],r0
 #define vsetq_lane_p16 vsetq_lane_u16
 
-float32x4_t vsetq_lane_f32(float32_t value, float32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
+_NEON2SSESTORAGE float32x4_t vsetq_lane_f32(float32_t value, float32x4_t vec, __constrange(0,3) int lane); // VMOV.32 d0[0],r0
 _NEON2SSE_INLINE float32x4_t vsetq_lane_f32(float32_t value, float32x4_t vec, __constrange(0,3) int lane)
 {
     float32_t val;
@@ -12096,7 +12098,7 @@ _NEON2SSE_INLINE float32x4_t vsetq_lane_f32(float32_t value, float32x4_t vec, __
     return vld1q_lane_f32(&val, vec,  lane);
 }
 
-int64x1_t vset_lane_s64(int64_t value, int64x1_t vec, __constrange(0,0) int lane); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int64x1_t vset_lane_s64(int64_t value, int64x1_t vec, __constrange(0,0) int lane); // VMOV d0,r0,r0
 _NEON2SSE_INLINE int64x1_t vset_lane_s64(int64_t value, int64x1_t vec, __constrange(0,0) int lane)
 {
     int64_t val;
@@ -12104,7 +12106,7 @@ _NEON2SSE_INLINE int64x1_t vset_lane_s64(int64_t value, int64x1_t vec, __constra
     return vld1_lane_s64(&val, vec,  lane);
 }
 
-uint64x1_t vset_lane_u64(uint64_t value, uint64x1_t vec, __constrange(0,0) int lane); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x1_t vset_lane_u64(uint64_t value, uint64x1_t vec, __constrange(0,0) int lane); // VMOV d0,r0,r0
 _NEON2SSE_INLINE uint64x1_t vset_lane_u64(uint64_t value, uint64x1_t vec, __constrange(0,0) int lane)
 {
     uint64_t val;
@@ -12112,7 +12114,7 @@ _NEON2SSE_INLINE uint64x1_t vset_lane_u64(uint64_t value, uint64x1_t vec, __cons
     return vld1_lane_u64(&val, vec,  lane);
 }
 
-int64x2_t vsetq_lane_s64(int64_t value, int64x2_t vec, __constrange(0,1) int lane); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int64x2_t vsetq_lane_s64(int64_t value, int64x2_t vec, __constrange(0,1) int lane); // VMOV d0,r0,r0
 _NEON2SSE_INLINE int64x2_t vsetq_lane_s64(int64_t value, int64x2_t vec, __constrange(0,1) int lane)
 {
     uint64_t val;
@@ -12120,60 +12122,60 @@ _NEON2SSE_INLINE int64x2_t vsetq_lane_s64(int64_t value, int64x2_t vec, __constr
     return vld1q_lane_s64(&val, vec,  lane);
 }
 
-uint64x2_t vsetq_lane_u64(uint64_t value, uint64x2_t vec, __constrange(0,1) int lane); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x2_t vsetq_lane_u64(uint64_t value, uint64x2_t vec, __constrange(0,1) int lane); // VMOV d0,r0,r0
 #define vsetq_lane_u64 vsetq_lane_s64
 
 // *******************************************************************************
 // **************** Initialize a vector from bit pattern ***************************
 // *******************************************************************************
 //These intrinsics create a vector from a literal bit pattern.
-int8x8_t vcreate_s8(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int8x8_t vcreate_s8(uint64_t a); // VMOV d0,r0,r0
 _NEON2SSE_INLINE int8x8_t vcreate_s8(uint64_t a)
 {
     return (*(__m64_128*)&(a)); //here we couldn't use macro due to possible immediate value usage
 }
 
-int16x4_t vcreate_s16(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int16x4_t vcreate_s16(uint64_t a); // VMOV d0,r0,r0
 #define vcreate_s16  vcreate_s8
 
-int32x2_t vcreate_s32(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int32x2_t vcreate_s32(uint64_t a); // VMOV d0,r0,r0
 #define vcreate_s32  vcreate_s8
 
-float16x4_t vcreate_f16(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE float16x4_t vcreate_f16(uint64_t a); // VMOV d0,r0,r0
 //no IA32 SIMD avalilable
 
-float32x2_t vcreate_f32(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE float32x2_t vcreate_f32(uint64_t a); // VMOV d0,r0,r0
 _NEON2SSE_INLINE float32x2_t vcreate_f32(uint64_t a)
 {
     return (*(__m64_128*)&(a)); //here we couldn't use macro due to possible immediate value usage
 }
 
-uint8x8_t vcreate_u8(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint8x8_t vcreate_u8(uint64_t a); // VMOV d0,r0,r0
 #define vcreate_u8 vcreate_s8
 
-uint16x4_t vcreate_u16(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint16x4_t vcreate_u16(uint64_t a); // VMOV d0,r0,r0
 #define vcreate_u16 vcreate_s16
 
-uint32x2_t vcreate_u32(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint32x2_t vcreate_u32(uint64_t a); // VMOV d0,r0,r0
 #define vcreate_u32 vcreate_s32
 
-uint64x1_t vcreate_u64(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x1_t vcreate_u64(uint64_t a); // VMOV d0,r0,r0
 #define vcreate_u64  vcreate_s8
 
 
-poly8x8_t vcreate_p8(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE poly8x8_t vcreate_p8(uint64_t a); // VMOV d0,r0,r0
 #define vcreate_p8 vcreate_u8
 
-poly16x4_t vcreate_p16(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE poly16x4_t vcreate_p16(uint64_t a); // VMOV d0,r0,r0
 #define vcreate_p16 vcreate_u16
 
-int64x1_t vcreate_s64(uint64_t a); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int64x1_t vcreate_s64(uint64_t a); // VMOV d0,r0,r0
 #define vcreate_s64 vcreate_u64
 
 //********************* Set all lanes to same value ********************************
 //*********************************************************************************
 //These intrinsics set all lanes to the same value.
-uint8x8_t   vdup_n_u8(uint8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE uint8x8_t   vdup_n_u8(uint8_t value); // VDUP.8 d0,r0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x8_t  vdup_n_u8(uint8_t value),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     uint8x8_t res;
@@ -12184,7 +12186,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint8x8_t  vdup_n_u8(uint8_t valu
     return res;
 }
 
-uint16x4_t   vdup_n_u16(uint16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE uint16x4_t   vdup_n_u16(uint16_t value); // VDUP.16 d0,r0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x4_t  vdup_n_u16(uint16_t value),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     uint16x4_t res;
@@ -12195,7 +12197,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint16x4_t  vdup_n_u16(uint16_t v
     return res;
 }
 
-uint32x2_t   vdup_n_u32(uint32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE uint32x2_t   vdup_n_u32(uint32_t value); // VDUP.32 d0,r0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t  vdup_n_u32(uint32_t value),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     uint32x2_t res;
@@ -12204,7 +12206,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(uint32x2_t  vdup_n_u32(uint32_t v
     return res;
 }
 
-int8x8_t   vdup_n_s8(int8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE int8x8_t   vdup_n_s8(int8_t value); // VDUP.8 d0,r0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x8_t  vdup_n_s8(int8_t value),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int8x8_t res;
@@ -12215,7 +12217,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x8_t  vdup_n_s8(int8_t value)
     return res;
 }
 
-int16x4_t   vdup_n_s16(int16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE int16x4_t   vdup_n_s16(int16_t value); // VDUP.16 d0,r0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x4_t  vdup_n_s16(int16_t value),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int16x4_t res;
@@ -12226,7 +12228,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int16x4_t  vdup_n_s16(int16_t val
     return res;
 }
 
-int32x2_t   vdup_n_s32(int32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE int32x2_t   vdup_n_s32(int32_t value); // VDUP.32 d0,r0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t  vdup_n_s32(int32_t value),  _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int32x2_t res;
@@ -12235,13 +12237,13 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t  vdup_n_s32(int32_t val
     return res;
 }
 
-poly8x8_t vdup_n_p8(poly8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE poly8x8_t vdup_n_p8(poly8_t value); // VDUP.8 d0,r0
 #define vdup_n_p8 vdup_n_u8
 
-poly16x4_t vdup_n_p16(poly16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE poly16x4_t vdup_n_p16(poly16_t value); // VDUP.16 d0,r0
 #define vdup_n_p16 vdup_n_s16
 
-float32x2_t vdup_n_f32(float32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE float32x2_t vdup_n_f32(float32_t value); // VDUP.32 d0,r0
 _NEON2SSE_INLINE float32x2_t vdup_n_f32(float32_t value)
 {
     float32x2_t res;
@@ -12250,34 +12252,34 @@ _NEON2SSE_INLINE float32x2_t vdup_n_f32(float32_t value)
     return res;
 }
 
-uint8x16_t   vdupq_n_u8(uint8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE uint8x16_t   vdupq_n_u8(uint8_t value); // VDUP.8 q0,r0
 #define vdupq_n_u8(value) _mm_set1_epi8((uint8_t) (value))
 
-uint16x8_t   vdupq_n_u16(uint16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE uint16x8_t   vdupq_n_u16(uint16_t value); // VDUP.16 q0,r0
 #define vdupq_n_u16(value) _mm_set1_epi16((uint16_t) (value))
 
-uint32x4_t   vdupq_n_u32(uint32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE uint32x4_t   vdupq_n_u32(uint32_t value); // VDUP.32 q0,r0
 #define vdupq_n_u32(value) _mm_set1_epi32((uint32_t) (value))
 
-int8x16_t   vdupq_n_s8(int8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE int8x16_t   vdupq_n_s8(int8_t value); // VDUP.8 q0,r0
 #define vdupq_n_s8 _mm_set1_epi8
 
-int16x8_t   vdupq_n_s16(int16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE int16x8_t   vdupq_n_s16(int16_t value); // VDUP.16 q0,r0
 #define vdupq_n_s16 _mm_set1_epi16
 
-int32x4_t   vdupq_n_s32(int32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE int32x4_t   vdupq_n_s32(int32_t value); // VDUP.32 q0,r0
 #define vdupq_n_s32 _mm_set1_epi32
 
-poly8x16_t vdupq_n_p8(poly8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE poly8x16_t vdupq_n_p8(poly8_t value); // VDUP.8 q0,r0
 #define  vdupq_n_p8 vdupq_n_u8
 
-poly16x8_t vdupq_n_p16(poly16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE poly16x8_t vdupq_n_p16(poly16_t value); // VDUP.16 q0,r0
 #define  vdupq_n_p16 vdupq_n_u16
 
-float32x4_t vdupq_n_f32(float32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE float32x4_t vdupq_n_f32(float32_t value); // VDUP.32 q0,r0
 #define vdupq_n_f32 _mm_set1_ps
 
-int64x1_t vdup_n_s64(int64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int64x1_t vdup_n_s64(int64_t value); // VMOV d0,r0,r0
 _NEON2SSE_INLINE int64x1_t vdup_n_s64(int64_t value)
 {
     int64x1_t res;
@@ -12285,7 +12287,7 @@ _NEON2SSE_INLINE int64x1_t vdup_n_s64(int64_t value)
     return res;
 }
 
-uint64x1_t vdup_n_u64(uint64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x1_t vdup_n_u64(uint64_t value); // VMOV d0,r0,r0
 _NEON2SSE_INLINE uint64x1_t  vdup_n_u64(uint64_t value)
 {
     uint64x1_t res;
@@ -12293,14 +12295,14 @@ _NEON2SSE_INLINE uint64x1_t  vdup_n_u64(uint64_t value)
     return res;
 }
 
-int64x2_t   vdupq_n_s64(int64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int64x2_t   vdupq_n_s64(int64_t value); // VMOV d0,r0,r0
 _NEON2SSE_INLINE int64x2_t   vdupq_n_s64(int64_t value)
 {
     _NEON2SSE_ALIGN_16 int64_t value2[2] = {value, value}; //value may be an immediate
     return LOAD_SI128(value2);
 }
 
-uint64x2_t   vdupq_n_u64(uint64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x2_t   vdupq_n_u64(uint64_t value); // VMOV d0,r0,r0
 _NEON2SSE_INLINE uint64x2_t   vdupq_n_u64(uint64_t value)
 {
     _NEON2SSE_ALIGN_16 uint64_t val[2] = {value, value}; //value may be an immediate
@@ -12310,76 +12312,76 @@ _NEON2SSE_INLINE uint64x2_t   vdupq_n_u64(uint64_t value)
 //****  Set all lanes to same value  ************************
 //Same functions as above - just aliaces.********************
 //Probably they reflect the fact that 128-bit functions versions use VMOV instruction **********
-uint8x8_t vmov_n_u8(uint8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE uint8x8_t vmov_n_u8(uint8_t value); // VDUP.8 d0,r0
 #define vmov_n_u8 vdup_n_s8
 
-uint16x4_t vmov_n_u16(uint16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE uint16x4_t vmov_n_u16(uint16_t value); // VDUP.16 d0,r0
 #define vmov_n_u16 vdup_n_s16
 
-uint32x2_t vmov_n_u32(uint32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE uint32x2_t vmov_n_u32(uint32_t value); // VDUP.32 d0,r0
 #define vmov_n_u32 vdup_n_u32
 
-int8x8_t vmov_n_s8(int8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE int8x8_t vmov_n_s8(int8_t value); // VDUP.8 d0,r0
 #define vmov_n_s8 vdup_n_s8
 
-int16x4_t vmov_n_s16(int16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE int16x4_t vmov_n_s16(int16_t value); // VDUP.16 d0,r0
 #define vmov_n_s16 vdup_n_s16
 
-int32x2_t vmov_n_s32(int32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE int32x2_t vmov_n_s32(int32_t value); // VDUP.32 d0,r0
 #define vmov_n_s32 vdup_n_s32
 
-poly8x8_t vmov_n_p8(poly8_t value); // VDUP.8 d0,r0
+_NEON2SSESTORAGE poly8x8_t vmov_n_p8(poly8_t value); // VDUP.8 d0,r0
 #define vmov_n_p8 vdup_n_u8
 
-poly16x4_t vmov_n_p16(poly16_t value); // VDUP.16 d0,r0
+_NEON2SSESTORAGE poly16x4_t vmov_n_p16(poly16_t value); // VDUP.16 d0,r0
 #define vmov_n_p16 vdup_n_s16
 
-float32x2_t vmov_n_f32(float32_t value); // VDUP.32 d0,r0
+_NEON2SSESTORAGE float32x2_t vmov_n_f32(float32_t value); // VDUP.32 d0,r0
 #define vmov_n_f32 vdup_n_f32
 
-uint8x16_t vmovq_n_u8(uint8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE uint8x16_t vmovq_n_u8(uint8_t value); // VDUP.8 q0,r0
 #define vmovq_n_u8 vdupq_n_u8
 
-uint16x8_t vmovq_n_u16(uint16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE uint16x8_t vmovq_n_u16(uint16_t value); // VDUP.16 q0,r0
 #define vmovq_n_u16 vdupq_n_s16
 
-uint32x4_t vmovq_n_u32(uint32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE uint32x4_t vmovq_n_u32(uint32_t value); // VDUP.32 q0,r0
 #define vmovq_n_u32 vdupq_n_u32
 
-int8x16_t vmovq_n_s8(int8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE int8x16_t vmovq_n_s8(int8_t value); // VDUP.8 q0,r0
 #define vmovq_n_s8 vdupq_n_s8
 
-int16x8_t vmovq_n_s16(int16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE int16x8_t vmovq_n_s16(int16_t value); // VDUP.16 q0,r0
 #define vmovq_n_s16 vdupq_n_s16
 
-int32x4_t vmovq_n_s32(int32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE int32x4_t vmovq_n_s32(int32_t value); // VDUP.32 q0,r0
 #define vmovq_n_s32 vdupq_n_s32
 
-poly8x16_t vmovq_n_p8(poly8_t value); // VDUP.8 q0,r0
+_NEON2SSESTORAGE poly8x16_t vmovq_n_p8(poly8_t value); // VDUP.8 q0,r0
 #define vmovq_n_p8 vdupq_n_u8
 
-poly16x8_t vmovq_n_p16(poly16_t value); // VDUP.16 q0,r0
+_NEON2SSESTORAGE poly16x8_t vmovq_n_p16(poly16_t value); // VDUP.16 q0,r0
 #define vmovq_n_p16 vdupq_n_s16
 
-float32x4_t vmovq_n_f32(float32_t value); // VDUP.32 q0,r0
+_NEON2SSESTORAGE float32x4_t vmovq_n_f32(float32_t value); // VDUP.32 q0,r0
 #define vmovq_n_f32 vdupq_n_f32
 
-int64x1_t vmov_n_s64(int64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int64x1_t vmov_n_s64(int64_t value); // VMOV d0,r0,r0
 #define vmov_n_s64 vdup_n_s64
 
-uint64x1_t vmov_n_u64(uint64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x1_t vmov_n_u64(uint64_t value); // VMOV d0,r0,r0
 #define vmov_n_u64 vdup_n_u64
 
-int64x2_t vmovq_n_s64(int64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE int64x2_t vmovq_n_s64(int64_t value); // VMOV d0,r0,r0
 #define vmovq_n_s64 vdupq_n_s64
 
-uint64x2_t vmovq_n_u64(uint64_t value); // VMOV d0,r0,r0
+_NEON2SSESTORAGE uint64x2_t vmovq_n_u64(uint64_t value); // VMOV d0,r0,r0
 #define vmovq_n_u64 vdupq_n_u64
 
 //**************Set all lanes to the value of one lane of a vector *************
 //****************************************************************************
 //here shuffle is better solution than lane extraction followed by set1 function
-uint8x8_t vdup_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
+_NEON2SSESTORAGE uint8x8_t vdup_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
 _NEON2SSE_INLINE uint8x8_t vdup_lane_u8(uint8x8_t vec, __constrange(0,7) int lane)
 {
     uint8x8_t res;
@@ -12392,7 +12394,7 @@ _NEON2SSE_INLINE uint8x8_t vdup_lane_u8(uint8x8_t vec, __constrange(0,7) int lan
     return res;
 }
 
-uint16x4_t vdup_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
+_NEON2SSESTORAGE uint16x4_t vdup_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
 _NEON2SSE_INLINE uint16x4_t vdup_lane_u16(uint16x4_t vec, __constrange(0,3) int lane)
 {
     uint16x4_t res;
@@ -12405,7 +12407,7 @@ _NEON2SSE_INLINE uint16x4_t vdup_lane_u16(uint16x4_t vec, __constrange(0,3) int
     return res;
 }
 
-uint32x2_t vdup_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
+_NEON2SSESTORAGE uint32x2_t vdup_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
 _NEON2SSE_INLINE uint32x2_t vdup_lane_u32(uint32x2_t vec, __constrange(0,1) int lane)
 {
     uint32x2_t res;
@@ -12414,22 +12416,22 @@ _NEON2SSE_INLINE uint32x2_t vdup_lane_u32(uint32x2_t vec, __constrange(0,1) int
     return res;
 }
 
-int8x8_t vdup_lane_s8(int8x8_t vec,  __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
+_NEON2SSESTORAGE int8x8_t vdup_lane_s8(int8x8_t vec,  __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
 #define vdup_lane_s8 vdup_lane_u8
 
-int16x4_t vdup_lane_s16(int16x4_t vec,  __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
+_NEON2SSESTORAGE int16x4_t vdup_lane_s16(int16x4_t vec,  __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
 #define vdup_lane_s16 vdup_lane_u16
 
-int32x2_t vdup_lane_s32(int32x2_t vec,  __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vdup_lane_s32(int32x2_t vec,  __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
 #define vdup_lane_s32 vdup_lane_u32
 
-poly8x8_t vdup_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
+_NEON2SSESTORAGE poly8x8_t vdup_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VDUP.8 d0,d0[0]
 #define vdup_lane_p8 vdup_lane_u8
 
-poly16x4_t vdup_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
+_NEON2SSESTORAGE poly16x4_t vdup_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VDUP.16 d0,d0[0]
 #define vdup_lane_p16 vdup_lane_s16
 
-float32x2_t vdup_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
+_NEON2SSESTORAGE float32x2_t vdup_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VDUP.32 d0,d0[0]
 _NEON2SSE_INLINE float32x2_t vdup_lane_f32(float32x2_t vec, __constrange(0,1) int lane)
 {
     float32x2_t res;
@@ -12438,7 +12440,7 @@ _NEON2SSE_INLINE float32x2_t vdup_lane_f32(float32x2_t vec, __constrange(0,1) in
     return res;
 }
 
-uint8x16_t vdupq_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
+_NEON2SSESTORAGE uint8x16_t vdupq_lane_u8(uint8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
 _NEON2SSE_INLINE uint8x16_t vdupq_lane_u8(uint8x8_t vec, __constrange(0,7) int lane) // VDUP.8 q0,d0[0]
 {
     const int8_t lane8 = (int8_t) lane;
@@ -12446,7 +12448,7 @@ _NEON2SSE_INLINE uint8x16_t vdupq_lane_u8(uint8x8_t vec, __constrange(0,7) int l
     return _mm_shuffle_epi8 (_pM128i(vec), *(__m128i*) lanemask8);
 }
 
-uint16x8_t vdupq_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
+_NEON2SSESTORAGE uint16x8_t vdupq_lane_u16(uint16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
 _NEON2SSE_INLINE uint16x8_t vdupq_lane_u16(uint16x4_t vec, __constrange(0,3) int lane) // VDUP.16 q0,d0[0]
 {
     //we could use 8bit shuffle for 16 bit as well
@@ -12457,41 +12459,41 @@ _NEON2SSE_INLINE uint16x8_t vdupq_lane_u16(uint16x4_t vec, __constrange(0,3) int
     return _mm_shuffle_epi8 (_pM128i(vec), *(__m128i*)lanemask_e16);
 }
 
-uint32x4_t vdupq_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
-_NEON2SSE_INLINE uint32x4_t vdupq_lane_u32(uint32x2_t vec, __constrange(0,1) int lane) 
+_NEON2SSESTORAGE uint32x4_t vdupq_lane_u32(uint32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
+_NEON2SSE_INLINE uint32x4_t vdupq_lane_u32(uint32x2_t vec, __constrange(0,1) int lane)
 {
     //need to use function not macro to make it gcc friendly and meet the immediate const requirement for _mm_shuffle_epi32
     if (lane == 1)
-        return _mm_shuffle_epi32 (_pM128i(vec), (1 | (1 << 2) | (1 << 4) | (1 << 6)) ); 
+        return _mm_shuffle_epi32 (_pM128i(vec), (1 | (1 << 2) | (1 << 4) | (1 << 6)) );
     else
-        return _mm_shuffle_epi32 (_pM128i(vec), 0); 
+        return _mm_shuffle_epi32 (_pM128i(vec), 0);
 }
 
-int8x16_t vdupq_lane_s8(int8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
+_NEON2SSESTORAGE int8x16_t vdupq_lane_s8(int8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
 #define vdupq_lane_s8 vdupq_lane_u8
 
-int16x8_t vdupq_lane_s16(int16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
+_NEON2SSESTORAGE int16x8_t vdupq_lane_s16(int16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
 #define vdupq_lane_s16 vdupq_lane_u16
 
-int32x4_t vdupq_lane_s32(int32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vdupq_lane_s32(int32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
 #define vdupq_lane_s32 vdupq_lane_u32
 
-poly8x16_t vdupq_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
+_NEON2SSESTORAGE poly8x16_t vdupq_lane_p8(poly8x8_t vec, __constrange(0,7) int lane); // VDUP.8 q0,d0[0]
 #define vdupq_lane_p8 vdupq_lane_u8
 
-poly16x8_t vdupq_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
+_NEON2SSESTORAGE poly16x8_t vdupq_lane_p16(poly16x4_t vec, __constrange(0,3) int lane); // VDUP.16 q0,d0[0]
 #define vdupq_lane_p16 vdupq_lane_s16
 
-float32x4_t vdupq_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
+_NEON2SSESTORAGE float32x4_t vdupq_lane_f32(float32x2_t vec, __constrange(0,1) int lane); // VDUP.32 q0,d0[0]
 #define  vdupq_lane_f32(vec, lane)  _mm_load1_ps((vec.m64_f32 + lane))
 
-int64x1_t vdup_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV d0,d0
+_NEON2SSESTORAGE int64x1_t vdup_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV d0,d0
 #define vdup_lane_s64(vec,lane) vec
 
-uint64x1_t vdup_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV d0,d0
+_NEON2SSESTORAGE uint64x1_t vdup_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV d0,d0
 #define vdup_lane_u64(vec,lane) vec
 
-int64x2_t vdupq_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV q0,q0
+_NEON2SSESTORAGE int64x2_t vdupq_lane_s64(int64x1_t vec, __constrange(0,0) int lane); // VMOV q0,q0
 _NEON2SSE_INLINE int64x2_t vdupq_lane_s64(int64x1_t vec, __constrange(0,0) int lane)
 {
     __m128i vec128;
@@ -12499,32 +12501,32 @@ _NEON2SSE_INLINE int64x2_t vdupq_lane_s64(int64x1_t vec, __constrange(0,0) int l
     return _mm_unpacklo_epi64(vec128,vec128);
 }
 
-uint64x2_t vdupq_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV q0,q0
+_NEON2SSESTORAGE uint64x2_t vdupq_lane_u64(uint64x1_t vec, __constrange(0,0) int lane); // VMOV q0,q0
 #define vdupq_lane_u64 vdupq_lane_s64
 
 // ********************************************************************
 // ********************  Combining vectors *****************************
 // ********************************************************************
 //These intrinsics join two 64 bit vectors into a single 128bit vector.
-int8x16_t   vcombine_s8(int8x8_t low, int8x8_t high); // VMOV d0,d0
+_NEON2SSESTORAGE int8x16_t   vcombine_s8(int8x8_t low, int8x8_t high); // VMOV d0,d0
 _NEON2SSE_INLINE int8x16_t  vcombine_s8(int8x8_t low, int8x8_t high)
 {
    return _mm_unpacklo_epi64 (_pM128i(low), _pM128i(high) );
 }
 
-int16x8_t   vcombine_s16(int16x4_t low, int16x4_t high); // VMOV d0,d0
+_NEON2SSESTORAGE int16x8_t   vcombine_s16(int16x4_t low, int16x4_t high); // VMOV d0,d0
 #define vcombine_s16 vcombine_s8
 
-int32x4_t   vcombine_s32(int32x2_t low, int32x2_t high); // VMOV d0,d0
+_NEON2SSESTORAGE int32x4_t   vcombine_s32(int32x2_t low, int32x2_t high); // VMOV d0,d0
 #define vcombine_s32 vcombine_s8
 
-int64x2_t   vcombine_s64(int64x1_t low, int64x1_t high); // VMOV d0,d0
+_NEON2SSESTORAGE int64x2_t   vcombine_s64(int64x1_t low, int64x1_t high); // VMOV d0,d0
 #define vcombine_s64 vcombine_s8
 
-float16x8_t vcombine_f16(float16x4_t low, float16x4_t high); // VMOV d0,d0
+_NEON2SSESTORAGE float16x8_t vcombine_f16(float16x4_t low, float16x4_t high); // VMOV d0,d0
 //current IA SIMD doesn't support float16
 
-float32x4_t vcombine_f32(float32x2_t low, float32x2_t high); // VMOV d0,d0
+_NEON2SSESTORAGE float32x4_t vcombine_f32(float32x2_t low, float32x2_t high); // VMOV d0,d0
 _NEON2SSE_INLINE float32x4_t vcombine_f32(float32x2_t low, float32x2_t high)
 {
     __m128i res;
@@ -12532,22 +12534,22 @@ _NEON2SSE_INLINE float32x4_t vcombine_f32(float32x2_t low, float32x2_t high)
     return _M128(res);
 }
 
-uint8x16_t   vcombine_u8(uint8x8_t low, uint8x8_t high); // VMOV d0,d0
+_NEON2SSESTORAGE uint8x16_t   vcombine_u8(uint8x8_t low, uint8x8_t high); // VMOV d0,d0
 #define vcombine_u8 vcombine_s8
 
-uint16x8_t   vcombine_u16(uint16x4_t low, uint16x4_t high); // VMOV d0,d0
+_NEON2SSESTORAGE uint16x8_t   vcombine_u16(uint16x4_t low, uint16x4_t high); // VMOV d0,d0
 #define vcombine_u16 vcombine_s16
 
-uint32x4_t   vcombine_u32(uint32x2_t low, uint32x2_t high); // VMOV d0,d0
+_NEON2SSESTORAGE uint32x4_t   vcombine_u32(uint32x2_t low, uint32x2_t high); // VMOV d0,d0
 #define vcombine_u32 vcombine_s32
 
-uint64x2_t   vcombine_u64(uint64x1_t low, uint64x1_t high); // VMOV d0,d0
+_NEON2SSESTORAGE uint64x2_t   vcombine_u64(uint64x1_t low, uint64x1_t high); // VMOV d0,d0
 #define vcombine_u64 vcombine_s64
 
-poly8x16_t   vcombine_p8(poly8x8_t low, poly8x8_t high); // VMOV d0,d0
+_NEON2SSESTORAGE poly8x16_t   vcombine_p8(poly8x8_t low, poly8x8_t high); // VMOV d0,d0
 #define vcombine_p8 vcombine_u8
 
-poly16x8_t   vcombine_p16(poly16x4_t low, poly16x4_t high); // VMOV d0,d0
+_NEON2SSESTORAGE poly16x8_t   vcombine_p16(poly16x4_t low, poly16x4_t high); // VMOV d0,d0
 #define vcombine_p16 vcombine_u16
 
 //**********************************************************************
@@ -12555,7 +12557,7 @@ poly16x8_t   vcombine_p16(poly16x4_t low, poly16x4_t high); // VMOV d0,d0
 //**********************************************************************
 //**************** Get high part ******************************************
 //These intrinsics split a 128 bit vector into 2 component 64 bit vectors
-int8x8_t vget_high_s8(int8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int8x8_t vget_high_s8(int8x16_t a); // VMOV d0,d0
 _NEON2SSE_INLINE int8x8_t vget_high_s8(int8x16_t a)
 {
     int8x8_t res64;
@@ -12564,7 +12566,7 @@ _NEON2SSE_INLINE int8x8_t vget_high_s8(int8x16_t a)
     return64(res);
 }
 
-int16x4_t vget_high_s16(int16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int16x4_t vget_high_s16(int16x8_t a); // VMOV d0,d0
 _NEON2SSE_INLINE int16x4_t vget_high_s16(int16x8_t a)
 {
     int16x4_t res64;
@@ -12573,7 +12575,7 @@ _NEON2SSE_INLINE int16x4_t vget_high_s16(int16x8_t a)
     return64(res);
 }
 
-int32x2_t vget_high_s32(int32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int32x2_t vget_high_s32(int32x4_t a); // VMOV d0,d0
 _NEON2SSE_INLINE int32x2_t vget_high_s32(int32x4_t a)
 {
     int32x2_t res64;
@@ -12582,7 +12584,7 @@ _NEON2SSE_INLINE int32x2_t vget_high_s32(int32x4_t a)
     return64(res);
 }
 
-int64x1_t vget_high_s64(int64x2_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int64x1_t vget_high_s64(int64x2_t a); // VMOV d0,d0
 _NEON2SSE_INLINE int64x1_t vget_high_s64(int64x2_t a)
 {
     int64x1_t res64;
@@ -12591,10 +12593,10 @@ _NEON2SSE_INLINE int64x1_t vget_high_s64(int64x2_t a)
     return64(res);
 }
 
-float16x4_t vget_high_f16(float16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE float16x4_t vget_high_f16(float16x8_t a); // VMOV d0,d0
 // IA32 SIMD doesn't work with 16bit floats currently
 
-float32x2_t vget_high_f32(float32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE float32x2_t vget_high_f32(float32x4_t a); // VMOV d0,d0
 _NEON2SSE_INLINE float32x2_t vget_high_f32(float32x4_t a)
 {
     __m128i res;
@@ -12603,58 +12605,58 @@ _NEON2SSE_INLINE float32x2_t vget_high_f32(float32x4_t a)
     return64(res);
 }
 
-uint8x8_t vget_high_u8(uint8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint8x8_t vget_high_u8(uint8x16_t a); // VMOV d0,d0
 #define vget_high_u8 vget_high_s8
 
-uint16x4_t vget_high_u16(uint16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint16x4_t vget_high_u16(uint16x8_t a); // VMOV d0,d0
 #define vget_high_u16 vget_high_s16
 
-uint32x2_t vget_high_u32(uint32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint32x2_t vget_high_u32(uint32x4_t a); // VMOV d0,d0
 #define vget_high_u32 vget_high_s32
 
-uint64x1_t vget_high_u64(uint64x2_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint64x1_t vget_high_u64(uint64x2_t a); // VMOV d0,d0
 #define vget_high_u64 vget_high_s64
 
-poly8x8_t vget_high_p8(poly8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE poly8x8_t vget_high_p8(poly8x16_t a); // VMOV d0,d0
 #define vget_high_p8 vget_high_u8
 
-poly16x4_t vget_high_p16(poly16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE poly16x4_t vget_high_p16(poly16x8_t a); // VMOV d0,d0
 #define vget_high_p16 vget_high_u16
 
 //********************** Get low part **********************
 //**********************************************************
-int8x8_t vget_low_s8(int8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int8x8_t vget_low_s8(int8x16_t a); // VMOV d0,d0
 _NEON2SSE_INLINE int8x8_t vget_low_s8(int8x16_t a) // VMOV d0,d0
 {
     int16x4_t res64;
     return64(a);
 }
 
-int16x4_t vget_low_s16(int16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int16x4_t vget_low_s16(int16x8_t a); // VMOV d0,d0
 _NEON2SSE_INLINE int16x4_t vget_low_s16(int16x8_t a) // VMOV d0,d0
 {
     int16x4_t res64;
     return64(a);
 }
 
-int32x2_t vget_low_s32(int32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int32x2_t vget_low_s32(int32x4_t a); // VMOV d0,d0
 _NEON2SSE_INLINE int32x2_t vget_low_s32(int32x4_t a) // VMOV d0,d0
 {
     int32x2_t res64;
     return64(a);
 }
 
-int64x1_t vget_low_s64(int64x2_t a); // VMOV d0,d0
+_NEON2SSESTORAGE int64x1_t vget_low_s64(int64x2_t a); // VMOV d0,d0
 _NEON2SSE_INLINE int64x1_t vget_low_s64(int64x2_t a) // VMOV d0,d0
 {
     int64x1_t res64;
     return64 (a);
 }
 
-float16x4_t vget_low_f16(float16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE float16x4_t vget_low_f16(float16x8_t a); // VMOV d0,d0
 // IA32 SIMD doesn't work with 16bit floats currently
 
-float32x2_t vget_low_f32(float32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE float32x2_t vget_low_f32(float32x4_t a); // VMOV d0,d0
 _NEON2SSE_INLINE float32x2_t vget_low_f32(float32x4_t a)
 {
     float32x2_t res64;
@@ -12662,22 +12664,22 @@ _NEON2SSE_INLINE float32x2_t vget_low_f32(float32x4_t a)
     return res64;
 }
 
-uint8x8_t vget_low_u8(uint8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint8x8_t vget_low_u8(uint8x16_t a); // VMOV d0,d0
 #define vget_low_u8 vget_low_s8
 
-uint16x4_t vget_low_u16(uint16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint16x4_t vget_low_u16(uint16x8_t a); // VMOV d0,d0
 #define vget_low_u16 vget_low_s16
 
-uint32x2_t vget_low_u32(uint32x4_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint32x2_t vget_low_u32(uint32x4_t a); // VMOV d0,d0
 #define vget_low_u32 vget_low_s32
 
-uint64x1_t vget_low_u64(uint64x2_t a); // VMOV d0,d0
+_NEON2SSESTORAGE uint64x1_t vget_low_u64(uint64x2_t a); // VMOV d0,d0
 #define vget_low_u64 vget_low_s64
 
-poly8x8_t vget_low_p8(poly8x16_t a); // VMOV d0,d0
+_NEON2SSESTORAGE poly8x8_t vget_low_p8(poly8x16_t a); // VMOV d0,d0
 #define vget_low_p8 vget_low_u8
 
-poly16x4_t vget_low_p16(poly16x8_t a); // VMOV d0,d0
+_NEON2SSESTORAGE poly16x4_t vget_low_p16(poly16x8_t a); // VMOV d0,d0
 #define vget_low_p16 vget_low_s16
 
 //**************************************************************************
@@ -12685,16 +12687,16 @@ poly16x4_t vget_low_p16(poly16x8_t a); // VMOV d0,d0
 //**************************************************************************
 //************* Convert from float ***************************************
 // need to set _MM_SET_ROUNDING_MODE ( x) accordingly
-int32x2_t   vcvt_s32_f32(float32x2_t a); // VCVT.S32.F32 d0, d0
+_NEON2SSESTORAGE int32x2_t   vcvt_s32_f32(float32x2_t a); // VCVT.S32.F32 d0, d0
 _NEON2SSE_INLINE int32x2_t   vcvt_s32_f32(float32x2_t a)
-{ 
+{
     int32x2_t res64;
     __m128i res;
     res =  _mm_cvtps_epi32(_pM128(a)); //use low 64 bits of result only
     return64(res);
 }
 
-uint32x2_t vcvt_u32_f32(float32x2_t a); // VCVT.U32.F32 d0, d0
+_NEON2SSESTORAGE uint32x2_t vcvt_u32_f32(float32x2_t a); // VCVT.U32.F32 d0, d0
 _NEON2SSE_INLINE uint32x2_t vcvt_u32_f32(float32x2_t a)
 {
     uint32x2_t res64;
@@ -12703,34 +12705,34 @@ _NEON2SSE_INLINE uint32x2_t vcvt_u32_f32(float32x2_t a)
     return64(res);
 }
 
-int32x4_t  vcvtq_s32_f32(float32x4_t a); // VCVT.S32.F32 q0, q0
+_NEON2SSESTORAGE int32x4_t  vcvtq_s32_f32(float32x4_t a); // VCVT.S32.F32 q0, q0
 _NEON2SSE_INLINE int32x4_t  vcvtq_s32_f32(float32x4_t a)
 {
     __m128 dif;
     __m128i res;
     //_mm_cvttps_epi32 incorrectly treats the case a > =2.14748364e+009, therefore the special processing is necessary
-    _NEON2SSE_ALIGN_16 float32_t fmax[] = { 2.14748364e+009, 2.14748364e+009, 2.14748364e+009, 2.14748364e+009 };
+    _NEON2SSE_ALIGN_16 static const float32_t fmax[] = { 2.14748364e+009f, 2.14748364e+009f, 2.14748364e+009f, 2.14748364e+009f };
     dif = _mm_cmpge_ps(a, *(__m128*)fmax);
     res = _mm_cvttps_epi32(a);
     return _mm_xor_si128(res, _M128i(dif));
 }
 
-uint32x4_t vcvtq_u32_f32(float32x4_t a); // VCVT.U32.F32 q0, q0
+_NEON2SSESTORAGE uint32x4_t vcvtq_u32_f32(float32x4_t a); // VCVT.U32.F32 q0, q0
 _NEON2SSE_INLINE uint32x4_t vcvtq_u32_f32(float32x4_t a) // VCVT.U32.F32 q0, q0
 {
     //No single instruction SSE solution  but we could implement it as following:
     __m128i res1, res2, zero, mask;
     __m128  max, min, dif;
-    _NEON2SSE_ALIGN_16 float32_t fmax[] = { 2.14748364e+009, 2.14748364e+009, 2.14748364e+009, 2.14748364e+009 };
-    _NEON2SSE_ALIGN_16 float32_t fmax_unsigned[] = { 4.29496729e+009, 4.29496729e+009, 4.29496729e+009, 4.29496729e+009 };
+    _NEON2SSE_ALIGN_16 static const float32_t fmax[] = { 2.14748364e+009f, 2.14748364e+009f, 2.14748364e+009f, 2.14748364e+009f };
+    _NEON2SSE_ALIGN_16 static const float32_t fmax_unsigned[] = { 4.29496729e+009f, 4.29496729e+009f, 4.29496729e+009f, 4.29496729e+009f };
     zero = _mm_setzero_si128();
     mask = _mm_cmpgt_epi32(_M128i(a), zero);
     min = _mm_and_ps(_M128(mask), a);
-    max = _mm_min_ps(min, *(__m128*)fmax_unsigned); //clamped in 0 - 4.29496729+009 
+    max = _mm_min_ps(min, *(__m128*)fmax_unsigned); //clamped in 0 - 4.29496729+009
 
     dif = _mm_sub_ps(max, *(__m128*)fmax);
-	mask = _mm_cmpgt_epi32(_M128i(dif),zero);
-	dif = _mm_and_ps(_M128(mask), dif);
+    mask = _mm_cmpgt_epi32(_M128i(dif),zero);
+    dif = _mm_and_ps(_M128(mask), dif);
 
     res1 = _mm_cvttps_epi32(dif);
     res2 = vcvtq_s32_f32(max);
@@ -12739,14 +12741,14 @@ _NEON2SSE_INLINE uint32x4_t vcvtq_u32_f32(float32x4_t a) // VCVT.U32.F32 q0, q0
 
 // ***** Convert to the fixed point  with   the number of fraction bits specified by b ***********
 //*************************************************************************************************
-int32x2_t vcvt_n_s32_f32(float32x2_t a, __constrange(1,32) int b); // VCVT.S32.F32 d0, d0, #32
+_NEON2SSESTORAGE int32x2_t vcvt_n_s32_f32(float32x2_t a, __constrange(1,32) int b); // VCVT.S32.F32 d0, d0, #32
 _NEON2SSE_INLINE int32x2_t vcvt_n_s32_f32(float32x2_t a, __constrange(1,32) int b)
 {
     int32x2_t res64;
     return64(vcvtq_n_s32_f32(_pM128(a),b));
 }
 
-uint32x2_t vcvt_n_u32_f32(float32x2_t a, __constrange(1,32) int b); // VCVT.U32.F32 d0, d0, #32
+_NEON2SSESTORAGE uint32x2_t vcvt_n_u32_f32(float32x2_t a, __constrange(1,32) int b); // VCVT.U32.F32 d0, d0, #32
 _NEON2SSE_INLINE uint32x2_t vcvt_n_u32_f32(float32x2_t a, __constrange(1,32) int b)
 {
     uint32x2_t res;
@@ -12757,11 +12759,11 @@ _NEON2SSE_INLINE uint32x2_t vcvt_n_u32_f32(float32x2_t a, __constrange(1,32) int
     return res;
 }
 
-int32x4_t vcvtq_n_s32_f32(float32x4_t a, __constrange(1,32) int b); // VCVT.S32.F32 q0, q0, #32
+_NEON2SSESTORAGE int32x4_t vcvtq_n_s32_f32(float32x4_t a, __constrange(1,32) int b); // VCVT.S32.F32 q0, q0, #32
 _NEON2SSE_INLINE int32x4_t vcvtq_n_s32_f32(float32x4_t a, __constrange(1,32) int b)
 {
     float convconst;
-    _NEON2SSE_ALIGN_16 uint32_t cmask[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
+    _NEON2SSE_ALIGN_16 static const uint32_t cmask[] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
     __m128 cconst128;
     __m128i mask, res;
     convconst = (float)(1 << b);
@@ -12771,7 +12773,7 @@ _NEON2SSE_INLINE int32x4_t vcvtq_n_s32_f32(float32x4_t a, __constrange(1,32) int
     return _mm_xor_si128 (res,  mask); //res saturated for 0x80000000
 }
 
-uint32x4_t vcvtq_n_u32_f32(float32x4_t a, __constrange(1,32) int b); // VCVT.U32.F32 q0, q0, #32
+_NEON2SSESTORAGE uint32x4_t vcvtq_n_u32_f32(float32x4_t a, __constrange(1,32) int b); // VCVT.U32.F32 q0, q0, #32
 _NEON2SSE_INLINE uint32x4_t vcvtq_n_u32_f32(float32x4_t a, __constrange(1,32) int b)
 {
     float convconst;
@@ -12782,7 +12784,7 @@ _NEON2SSE_INLINE uint32x4_t vcvtq_n_u32_f32(float32x4_t a, __constrange(1,32) in
 }
 
 
-int32x4_t vcvtnq_s32_f32(float32x4_t a); // VCVTN.S32.F32 q0, q0
+_NEON2SSESTORAGE int32x4_t vcvtnq_s32_f32(float32x4_t a); // VCVTN.S32.F32 q0, q0
 _NEON2SSE_INLINE int32x4_t vcvtnq_s32_f32(float32x4_t a)
 {
   return _mm_cvtps_epi32(a);
@@ -12790,7 +12792,7 @@ _NEON2SSE_INLINE int32x4_t vcvtnq_s32_f32(float32x4_t a)
 
 //***************** Convert to float *************************
 //*************************************************************
-float32x2_t vcvt_f32_s32(int32x2_t a); // VCVT.F32.S32 d0, d0
+_NEON2SSESTORAGE float32x2_t vcvt_f32_s32(int32x2_t a); // VCVT.F32.S32 d0, d0
 _NEON2SSE_INLINE float32x2_t vcvt_f32_s32(int32x2_t a) //use low 64 bits
 {
     float32x2_t res;
@@ -12799,7 +12801,7 @@ _NEON2SSE_INLINE float32x2_t vcvt_f32_s32(int32x2_t a) //use low 64 bits
     return res;
 }
 
-float32x2_t vcvt_f32_u32(uint32x2_t a); // VCVT.F32.U32 d0, d0
+_NEON2SSESTORAGE float32x2_t vcvt_f32_u32(uint32x2_t a); // VCVT.F32.U32 d0, d0
 _NEON2SSE_INLINE float32x2_t vcvt_f32_u32(uint32x2_t a)
 {
     float32x2_t res;
@@ -12808,10 +12810,10 @@ _NEON2SSE_INLINE float32x2_t vcvt_f32_u32(uint32x2_t a)
     return res;
 }
 
-float32x4_t vcvtq_f32_s32(int32x4_t a); // VCVT.F32.S32 q0, q0
+_NEON2SSESTORAGE float32x4_t vcvtq_f32_s32(int32x4_t a); // VCVT.F32.S32 q0, q0
 #define vcvtq_f32_s32(a) _mm_cvtepi32_ps(a)
 
-float32x4_t vcvtq_f32_u32(uint32x4_t a); // VCVT.F32.U32 q0, q0
+_NEON2SSESTORAGE float32x4_t vcvtq_f32_u32(uint32x4_t a); // VCVT.F32.U32 q0, q0
 _NEON2SSE_INLINE float32x4_t vcvtq_f32_u32(uint32x4_t a) // VCVT.F32.U32 q0, q0
 {
     //solution may be not optimal
@@ -12829,7 +12831,7 @@ _NEON2SSE_INLINE float32x4_t vcvtq_f32_u32(uint32x4_t a) // VCVT.F32.U32 q0, q0
 }
 
 // ***** Convert to the float from fixed point  with   the number of fraction bits specified by b ***********
-float32x2_t vcvt_n_f32_s32(int32x2_t a, __constrange(1,32) int b); // VCVT.F32.S32 d0, d0, #32
+_NEON2SSESTORAGE float32x2_t vcvt_n_f32_s32(int32x2_t a, __constrange(1,32) int b); // VCVT.F32.S32 d0, d0, #32
 _NEON2SSE_INLINE float32x2_t vcvt_n_f32_s32(int32x2_t a, __constrange(1,32) int b)
 {
     float32x2_t res;
@@ -12840,7 +12842,7 @@ _NEON2SSE_INLINE float32x2_t vcvt_n_f32_s32(int32x2_t a, __constrange(1,32) int
     return res;
 }
 
-float32x2_t vcvt_n_f32_u32(uint32x2_t a, __constrange(1,32) int b); // VCVT.F32.U32 d0, d0, #32
+_NEON2SSESTORAGE float32x2_t vcvt_n_f32_u32(uint32x2_t a, __constrange(1,32) int b); // VCVT.F32.U32 d0, d0, #32
 _NEON2SSE_INLINE float32x2_t vcvt_n_f32_u32(uint32x2_t a, __constrange(1,32) int b) // VCVT.F32.U32 d0, d0, #32
 {
     float32x2_t res;
@@ -12851,7 +12853,7 @@ _NEON2SSE_INLINE float32x2_t vcvt_n_f32_u32(uint32x2_t a, __constrange(1,32) int
     return res;
 }
 
-float32x4_t vcvtq_n_f32_s32(int32x4_t a, __constrange(1,32) int b); // VCVT.F32.S32 q0, q0, #32
+_NEON2SSESTORAGE float32x4_t vcvtq_n_f32_s32(int32x4_t a, __constrange(1,32) int b); // VCVT.F32.S32 q0, q0, #32
 _NEON2SSE_INLINE float32x4_t vcvtq_n_f32_s32(int32x4_t a, __constrange(1,32) int b)
 {
     float convconst;
@@ -12862,7 +12864,7 @@ _NEON2SSE_INLINE float32x4_t vcvtq_n_f32_s32(int32x4_t a, __constrange(1,32) int
     return _mm_mul_ps(af,cconst128);
 }
 
-float32x4_t vcvtq_n_f32_u32(uint32x4_t a, __constrange(1,32) int b); // VCVT.F32.U32 q0, q0, #32
+_NEON2SSESTORAGE float32x4_t vcvtq_n_f32_u32(uint32x4_t a, __constrange(1,32) int b); // VCVT.F32.U32 q0, q0, #32
 _NEON2SSE_INLINE float32x4_t vcvtq_n_f32_u32(uint32x4_t a, __constrange(1,32) int b)
 {
     float convconst;
@@ -12875,15 +12877,15 @@ _NEON2SSE_INLINE float32x4_t vcvtq_n_f32_u32(uint32x4_t a, __constrange(1,32) in
 
 //**************Convert between floats ***********************
 //************************************************************
-float16x4_t vcvt_f16_f32(float32x4_t a); // VCVT.F16.F32 d0, q0
+_NEON2SSESTORAGE float16x4_t vcvt_f16_f32(float32x4_t a); // VCVT.F16.F32 d0, q0
 //Intel SIMD doesn't support 16bits floats curently
 
-float32x4_t vcvt_f32_f16(float16x4_t a); // VCVT.F32.F16 q0, d0
+_NEON2SSESTORAGE float32x4_t vcvt_f32_f16(float16x4_t a); // VCVT.F32.F16 q0, d0
 //Intel SIMD doesn't support 16bits floats curently, the only solution is to store 16bit floats and load as 32 bits
 
 //************Vector narrow integer conversion (truncation) ******************
 //****************************************************************************
-int8x8_t vmovn_s16(int16x8_t a); // VMOVN.I16 d0,q0
+_NEON2SSESTORAGE int8x8_t vmovn_s16(int16x8_t a); // VMOVN.I16 d0,q0
 _NEON2SSE_INLINE int8x8_t vmovn_s16(int16x8_t a) // VMOVN.I16 d0,q0
 {
     int8x8_t res64;
@@ -12892,7 +12894,7 @@ _NEON2SSE_INLINE int8x8_t vmovn_s16(int16x8_t a) // VMOVN.I16 d0,q0
     return64(res);
 }
 
-int16x4_t vmovn_s32(int32x4_t a); // VMOVN.I32 d0,q0
+_NEON2SSESTORAGE int16x4_t vmovn_s32(int32x4_t a); // VMOVN.I32 d0,q0
 _NEON2SSE_INLINE int16x4_t vmovn_s32(int32x4_t a) // VMOVN.I32 d0,q0
 {
     int16x4_t res64;
@@ -12901,7 +12903,7 @@ _NEON2SSE_INLINE int16x4_t vmovn_s32(int32x4_t a) // VMOVN.I32 d0,q0
     return64(res);
 }
 
-int32x2_t vmovn_s64(int64x2_t a); // VMOVN.I64 d0,q0
+_NEON2SSESTORAGE int32x2_t vmovn_s64(int64x2_t a); // VMOVN.I64 d0,q0
 _NEON2SSE_INLINE int32x2_t vmovn_s64(int64x2_t a)
 {
     //may be not effective compared with a serial implementation
@@ -12911,56 +12913,56 @@ _NEON2SSE_INLINE int32x2_t vmovn_s64(int64x2_t a)
     return64(res);
 }
 
-uint8x8_t vmovn_u16(uint16x8_t a); // VMOVN.I16 d0,q0
+_NEON2SSESTORAGE uint8x8_t vmovn_u16(uint16x8_t a); // VMOVN.I16 d0,q0
 #define vmovn_u16 vmovn_s16
 
-uint16x4_t vmovn_u32(uint32x4_t a); // VMOVN.I32 d0,q0
+_NEON2SSESTORAGE uint16x4_t vmovn_u32(uint32x4_t a); // VMOVN.I32 d0,q0
 #define vmovn_u32 vmovn_s32
 
-uint32x2_t vmovn_u64(uint64x2_t a); // VMOVN.I64 d0,q0
+_NEON2SSESTORAGE uint32x2_t vmovn_u64(uint64x2_t a); // VMOVN.I64 d0,q0
 #define vmovn_u64 vmovn_s64
 
 //**************** Vector long move   ***********************
 //***********************************************************
-int16x8_t vmovl_s8(int8x8_t a); // VMOVL.S8 q0,d0
+_NEON2SSESTORAGE int16x8_t vmovl_s8(int8x8_t a); // VMOVL.S8 q0,d0
 _NEON2SSE_INLINE int16x8_t vmovl_s8(int8x8_t a)
 {
     return _MM_CVTEPI8_EPI16(_pM128i(a)); //SSE4.1
 }
 
-int32x4_t vmovl_s16(int16x4_t a); // VMOVL.S16 q0,d0
-_NEON2SSE_INLINE int32x4_t vmovl_s16(int16x4_t a) 
+_NEON2SSESTORAGE int32x4_t vmovl_s16(int16x4_t a); // VMOVL.S16 q0,d0
+_NEON2SSE_INLINE int32x4_t vmovl_s16(int16x4_t a)
 {
     return _MM_CVTEPI16_EPI32(_pM128i(a)); //SSE4.1
 }
 
-int64x2_t vmovl_s32(int32x2_t a); // VMOVL.S32 q0,d0
-_NEON2SSE_INLINE int64x2_t  vmovl_s32(int32x2_t a)  
+_NEON2SSESTORAGE int64x2_t vmovl_s32(int32x2_t a); // VMOVL.S32 q0,d0
+_NEON2SSE_INLINE int64x2_t  vmovl_s32(int32x2_t a)
 {
     return _MM_CVTEPI32_EPI64(_pM128i(a)); //SSE4.1
 }
 
-uint16x8_t vmovl_u8(uint8x8_t a); // VMOVL.U8 q0,d0
-_NEON2SSE_INLINE uint16x8_t vmovl_u8(uint8x8_t a) 
+_NEON2SSESTORAGE uint16x8_t vmovl_u8(uint8x8_t a); // VMOVL.U8 q0,d0
+_NEON2SSE_INLINE uint16x8_t vmovl_u8(uint8x8_t a)
 {
     return _MM_CVTEPU8_EPI16(_pM128i(a)); //SSE4.1
 }
 
-uint32x4_t vmovl_u16(uint16x4_t a); // VMOVL.s16 q0,d0
+_NEON2SSESTORAGE uint32x4_t vmovl_u16(uint16x4_t a); // VMOVL.s16 q0,d0
 _NEON2SSE_INLINE uint32x4_t  vmovl_u16(uint16x4_t a)
 {
     return _MM_CVTEPU16_EPI32(_pM128i(a)); //SSE4.1
 }
 
-uint64x2_t vmovl_u32(uint32x2_t a); // VMOVL.U32 q0,d0
-_NEON2SSE_INLINE uint64x2_t  vmovl_u32(uint32x2_t a) 
+_NEON2SSESTORAGE uint64x2_t vmovl_u32(uint32x2_t a); // VMOVL.U32 q0,d0
+_NEON2SSE_INLINE uint64x2_t  vmovl_u32(uint32x2_t a)
 {
     return _MM_CVTEPU32_EPI64(_pM128i(a)); //SSE4.1
 }
 
 //*************Vector saturating narrow integer*****************
 //**************************************************************
-int8x8_t   vqmovn_s16(int16x8_t a); // VQMOVN.S16 d0,q0
+_NEON2SSESTORAGE int8x8_t   vqmovn_s16(int16x8_t a); // VQMOVN.S16 d0,q0
 _NEON2SSE_INLINE int8x8_t   vqmovn_s16(int16x8_t a)
 {
     int8x8_t res64;
@@ -12969,7 +12971,7 @@ _NEON2SSE_INLINE int8x8_t   vqmovn_s16(int16x8_t a)
     return64(res);
 }
 
-int16x4_t   vqmovn_s32(int32x4_t a); // VQMOVN.S32 d0,q0
+_NEON2SSESTORAGE int16x4_t   vqmovn_s32(int32x4_t a); // VQMOVN.S32 d0,q0
 _NEON2SSE_INLINE int16x4_t   vqmovn_s32(int32x4_t a)
 {
     int16x4_t res64;
@@ -12978,7 +12980,7 @@ _NEON2SSE_INLINE int16x4_t   vqmovn_s32(int32x4_t a)
     return64(res);
 }
 
-int32x2_t vqmovn_s64(int64x2_t a); // VQMOVN.S64 d0,q0
+_NEON2SSESTORAGE int32x2_t vqmovn_s64(int64x2_t a); // VQMOVN.S64 d0,q0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqmovn_s64(int64x2_t a),_NEON2SSE_REASON_SLOW_SERIAL) //no effective SIMD solution
 {
     int32x2_t res;
@@ -12993,7 +12995,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqmovn_s64(int64x2_t a)
     return res;
 }
 
-uint8x8_t vqmovn_u16(uint16x8_t a); // VQMOVN.s16 d0,q0
+_NEON2SSESTORAGE uint8x8_t vqmovn_u16(uint16x8_t a); // VQMOVN.s16 d0,q0
 _NEON2SSE_INLINE uint8x8_t vqmovn_u16(uint16x8_t a) // VQMOVN.s16 d0,q0
 {
     //no uint16 to uint8 conversion in SSE, need truncate to max signed first. Also trying to avoid _mm_shuffle_epi8 because of its big latency for old Atom CPUs
@@ -13008,7 +13010,7 @@ _NEON2SSE_INLINE uint8x8_t vqmovn_u16(uint16x8_t a) // VQMOVN.s16 d0,q0
     return64(a_trunc);
 }
 
-uint16x4_t vqmovn_u32(uint32x4_t a); // VQMOVN.U32 d0,q0
+_NEON2SSESTORAGE uint16x4_t vqmovn_u32(uint32x4_t a); // VQMOVN.U32 d0,q0
 _NEON2SSE_INLINE uint16x4_t vqmovn_u32(uint32x4_t a) // VQMOVN.U32 d0,q0
 {
      #ifdef USE_SSE4
@@ -13031,12 +13033,12 @@ _NEON2SSE_INLINE uint16x4_t vqmovn_u32(uint32x4_t a) // VQMOVN.U32 d0,q0
         mask = _mm_cmpeq_epi16(mask,mask); //all fff
         mask = _mm_andnot_si128(res_hi,mask); //inverst res_hi to get >16 bits numbers
         res_hi = _mm_or_si128(a, mask); //saturated res
-        res_hi = _mm_shuffle_epi8 (res_hi, *(__m128i*) mask8_32_even_odd); //go to 16 bits 
-        return64(res_hi);  
+        res_hi = _mm_shuffle_epi8 (res_hi, *(__m128i*) mask8_32_even_odd); //go to 16 bits
+        return64(res_hi);
     #endif
 }
 
-uint32x2_t vqmovn_u64(uint64x2_t a); // VQMOVN.U64 d0,q0
+_NEON2SSESTORAGE uint32x2_t vqmovn_u64(uint64x2_t a); // VQMOVN.U64 d0,q0
 _NEON2SSE_INLINE uint32x2_t vqmovn_u64(uint64x2_t a)
 {
     //serial solution may be faster
@@ -13053,7 +13055,7 @@ _NEON2SSE_INLINE uint32x2_t vqmovn_u64(uint64x2_t a)
 }
 //************* Vector saturating narrow integer signed->unsigned **************
 //*****************************************************************************
-uint8x8_t vqmovun_s16(int16x8_t a); // VQMOVUN.S16 d0,q0
+_NEON2SSESTORAGE uint8x8_t vqmovun_s16(int16x8_t a); // VQMOVUN.S16 d0,q0
 _NEON2SSE_INLINE uint8x8_t vqmovun_s16(int16x8_t a)
 {
     uint8x8_t res64;
@@ -13062,7 +13064,7 @@ _NEON2SSE_INLINE uint8x8_t vqmovun_s16(int16x8_t a)
     return64(res);
 }
 
-uint16x4_t vqmovun_s32(int32x4_t a); // VQMOVUN.S32 d0,q0
+_NEON2SSESTORAGE uint16x4_t vqmovun_s32(int32x4_t a); // VQMOVUN.S32 d0,q0
 _NEON2SSE_INLINE uint16x4_t vqmovun_s32(int32x4_t a)
 {
     uint16x4_t res64;
@@ -13071,7 +13073,7 @@ _NEON2SSE_INLINE uint16x4_t vqmovun_s32(int32x4_t a)
     return64(res);
 }
 
-uint32x2_t vqmovun_s64(int64x2_t a); // VQMOVUN.S64 d0,q0
+_NEON2SSESTORAGE uint32x2_t vqmovun_s64(int64x2_t a); // VQMOVUN.S64 d0,q0
 _NEON2SSE_INLINE uint32x2_t vqmovun_s64(int64x2_t a)
 {
     uint32x2_t res64;
@@ -13092,7 +13094,7 @@ _NEON2SSE_INLINE uint32x2_t vqmovun_s64(int64x2_t a)
 //VTBL (Vector Table Lookup) uses byte indexes in a control vector to look up byte values
 //in a table and generate a new vector. Indexes out of range return 0.
 //for Intel SIMD we need to set the MSB to 1 for zero return
-uint8x8_t vtbl1_u8(uint8x8_t a, uint8x8_t b); // VTBL.8 d0, {d0}, d0
+_NEON2SSESTORAGE uint8x8_t vtbl1_u8(uint8x8_t a, uint8x8_t b); // VTBL.8 d0, {d0}, d0
 _NEON2SSE_INLINE uint8x8_t vtbl1_u8(uint8x8_t a, uint8x8_t b)
 {
     uint8x8_t res64;
@@ -13105,13 +13107,13 @@ _NEON2SSE_INLINE uint8x8_t vtbl1_u8(uint8x8_t a, uint8x8_t b)
     return64(bmask);
 }
 
-int8x8_t vtbl1_s8(int8x8_t a,  int8x8_t b); // VTBL.8 d0, {d0}, d0
+_NEON2SSESTORAGE int8x8_t vtbl1_s8(int8x8_t a,  int8x8_t b); // VTBL.8 d0, {d0}, d0
 #define vtbl1_s8 vtbl1_u8
 
-poly8x8_t vtbl1_p8(poly8x8_t a, uint8x8_t b); // VTBL.8 d0, {d0}, d0
+_NEON2SSESTORAGE poly8x8_t vtbl1_p8(poly8x8_t a, uint8x8_t b); // VTBL.8 d0, {d0}, d0
 #define vtbl1_p8 vtbl1_u8
 
-uint8x8_t vtbl2_u8(uint8x8x2_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1}, d0
+_NEON2SSESTORAGE uint8x8_t vtbl2_u8(uint8x8x2_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1}, d0
 _NEON2SSE_INLINE uint8x8_t vtbl2_u8(uint8x8x2_t a, uint8x8_t b)
 {
     uint8x8_t res64;
@@ -13131,7 +13133,7 @@ _NEON2SSE_INLINE uint8x8_t vtbl2_u8(uint8x8x2_t a, uint8x8_t b)
 //poly8x8_t vtbl2_p8(poly8x8x2_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1}, d0
 #define vtbl2_p8 vtbl2_u8
 
-uint8x8_t vtbl3_u8(uint8x8x3_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1, d2}, d0
+_NEON2SSESTORAGE uint8x8_t vtbl3_u8(uint8x8x3_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1, d2}, d0
 _NEON2SSE_INLINE uint8x8_t vtbl3_u8(uint8x8x3_t a, uint8x8_t b)
 {
     //solution may be not optimal
@@ -13150,13 +13152,13 @@ _NEON2SSE_INLINE uint8x8_t vtbl3_u8(uint8x8x3_t a, uint8x8_t b)
     return64(sh0);
 }
 
-int8x8_t vtbl3_s8(int8x8x3_t a, int8x8_t b); // VTBL.8 d0, {d0, d1, d2}, d0
+_NEON2SSESTORAGE int8x8_t vtbl3_s8(int8x8x3_t a, int8x8_t b); // VTBL.8 d0, {d0, d1, d2}, d0
 #define vtbl3_s8 vtbl3_u8
 
-poly8x8_t vtbl3_p8(poly8x8x3_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1, d2}, d0
+_NEON2SSESTORAGE poly8x8_t vtbl3_p8(poly8x8x3_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1, d2}, d0
 #define vtbl3_p8 vtbl3_u8
 
-uint8x8_t vtbl4_u8(uint8x8x4_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1, d2, d3}, d0
+_NEON2SSESTORAGE uint8x8_t vtbl4_u8(uint8x8x4_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1, d2, d3}, d0
 _NEON2SSE_INLINE uint8x8_t vtbl4_u8(uint8x8x4_t a, uint8x8_t b)
 {
     //solution may be not optimal
@@ -13176,10 +13178,10 @@ _NEON2SSE_INLINE uint8x8_t vtbl4_u8(uint8x8x4_t a, uint8x8_t b)
     return64(sh0);
 }
 
-int8x8_t vtbl4_s8(int8x8x4_t a, int8x8_t b); // VTBL.8 d0, {d0, d1, d2, d3}, d0
+_NEON2SSESTORAGE int8x8_t vtbl4_s8(int8x8x4_t a, int8x8_t b); // VTBL.8 d0, {d0, d1, d2, d3}, d0
 #define vtbl4_s8 vtbl4_u8
 
-poly8x8_t vtbl4_p8(poly8x8x4_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1, d2, d3}, d0
+_NEON2SSESTORAGE poly8x8_t vtbl4_p8(poly8x8x4_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1, d2, d3}, d0
 #define vtbl4_p8 vtbl4_u8
 
 //****************** Extended table look up intrinsics ***************************
@@ -13187,7 +13189,7 @@ poly8x8_t vtbl4_p8(poly8x8x4_t a, uint8x8_t b); // VTBL.8 d0, {d0, d1, d2, d3},
 //VTBX (Vector Table Extension) works in the same way as VTBL do,
 // except that indexes out of range leave the destination element unchanged.
 
-uint8x8_t vtbx1_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VTBX.8 d0, {d0}, d0
+_NEON2SSESTORAGE uint8x8_t vtbx1_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c); // VTBX.8 d0, {d0}, d0
 _NEON2SSE_INLINE uint8x8_t vtbx1_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c)
 {
     uint8x8_t res64;
@@ -13202,13 +13204,13 @@ _NEON2SSE_INLINE uint8x8_t vtbx1_u8(uint8x8_t a, uint8x8_t b, uint8x8_t c)
     return64(sh);
 }
 
-int8x8_t vtbx1_s8(int8x8_t a,  int8x8_t b, int8x8_t c); // VTBX.8 d0, {d0}, d0
+_NEON2SSESTORAGE int8x8_t vtbx1_s8(int8x8_t a,  int8x8_t b, int8x8_t c); // VTBX.8 d0, {d0}, d0
 #define vtbx1_s8 vtbx1_u8
 
-poly8x8_t vtbx1_p8(poly8x8_t a, poly8x8_t b, uint8x8_t c); // VTBX.8 d0, {d0}, d0
+_NEON2SSESTORAGE poly8x8_t vtbx1_p8(poly8x8_t a, poly8x8_t b, uint8x8_t c); // VTBX.8 d0, {d0}, d0
 #define vtbx1_p8 vtbx1_u8
 
-uint8x8_t vtbx2_u8(uint8x8_t a, uint8x8x2_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1}, d0
+_NEON2SSESTORAGE uint8x8_t vtbx2_u8(uint8x8_t a, uint8x8x2_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1}, d0
 _NEON2SSE_INLINE uint8x8_t vtbx2_u8(uint8x8_t a, uint8x8x2_t b, uint8x8_t c)
 {
     uint8x8_t res64;
@@ -13230,7 +13232,7 @@ _NEON2SSE_INLINE uint8x8_t vtbx2_u8(uint8x8_t a, uint8x8x2_t b, uint8x8_t c)
 //poly8x8_t vtbx2_p8(poly8x8_t a, poly8x8x2_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1}, d0
 #define vtbx2_p8 vtbx2_u8
 
-uint8x8_t vtbx3_u8(uint8x8_t a, uint8x8x3_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
+_NEON2SSESTORAGE uint8x8_t vtbx3_u8(uint8x8_t a, uint8x8x3_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
 _NEON2SSE_INLINE uint8x8_t vtbx3_u8(uint8x8_t a, uint8x8x3_t b, uint8x8_t c)
 {
     //solution may be not optimal
@@ -13251,13 +13253,13 @@ _NEON2SSE_INLINE uint8x8_t vtbx3_u8(uint8x8_t a, uint8x8x3_t b, uint8x8_t c)
     return64(sh0);
 }
 
-int8x8_t vtbx3_s8(int8x8_t a, int8x8x3_t b, int8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
+_NEON2SSESTORAGE int8x8_t vtbx3_s8(int8x8_t a, int8x8x3_t b, int8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
 #define vtbx3_s8 vtbx3_u8
 
-poly8x8_t vtbx3_p8(poly8x8_t a, poly8x8x3_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
+_NEON2SSESTORAGE poly8x8_t vtbx3_p8(poly8x8_t a, poly8x8x3_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2}, d0
 #define vtbx3_p8 vtbx3_u8
 
-uint8x8_t vtbx4_u8(uint8x8_t a, uint8x8x4_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
+_NEON2SSESTORAGE uint8x8_t vtbx4_u8(uint8x8_t a, uint8x8x4_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
 _NEON2SSE_INLINE uint8x8_t vtbx4_u8(uint8x8_t a, uint8x8x4_t b, uint8x8_t c)
 {
     //solution may be not optimal
@@ -13280,10 +13282,10 @@ _NEON2SSE_INLINE uint8x8_t vtbx4_u8(uint8x8_t a, uint8x8x4_t b, uint8x8_t c)
     return64(sh0);
 }
 
-int8x8_t vtbx4_s8(int8x8_t a, int8x8x4_t b, int8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
+_NEON2SSESTORAGE int8x8_t vtbx4_s8(int8x8_t a, int8x8x4_t b, int8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
 #define vtbx4_s8 vtbx4_u8
 
-poly8x8_t vtbx4_p8(poly8x8_t a, poly8x8x4_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
+_NEON2SSESTORAGE poly8x8_t vtbx4_p8(poly8x8_t a, poly8x8x4_t b, uint8x8_t c); // VTBX.8 d0, {d0, d1, d2, d3}, d0
 #define vtbx4_p8 vtbx4_u8
 
 //*************************************************************************************************
@@ -13292,7 +13294,7 @@ poly8x8_t vtbx4_p8(poly8x8_t a, poly8x8x4_t b, uint8x8_t c); // VTBX.8 d0, {d0,
 
 //******* Vector multiply accumulate by scalar *************************************************
 //**********************************************************************************************
-int16x4_t vmla_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLA.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE int16x4_t vmla_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLA.I16 d0, d0, d0[0]
 _NEON2SSE_INLINE int16x4_t vmla_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l) // VMLA.I16 d0, d0, d0[0]
 {
     int16_t c;
@@ -13302,7 +13304,7 @@ _NEON2SSE_INLINE int16x4_t vmla_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v,
     return vmla_s16(a, b, scalar);
 }
 
-int32x2_t vmla_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLA.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE int32x2_t vmla_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLA.I32 d0, d0, d0[0]
 _NEON2SSE_INLINE int32x2_t vmla_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l) // VMLA.I32 d0, d0, d0[0]
 {
     int32_t c;
@@ -13312,14 +13314,14 @@ _NEON2SSE_INLINE int32x2_t vmla_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v,
     return vmla_s32(a, b, scalar);
 }
 
-uint16x4_t vmla_lane_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLA.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE uint16x4_t vmla_lane_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLA.I16 d0, d0, d0[0]
 #define vmla_lane_u16 vmla_lane_s16
 
 
-uint32x2_t vmla_lane_u32(uint32x2_t a,  uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLA.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE uint32x2_t vmla_lane_u32(uint32x2_t a,  uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLA.I32 d0, d0, d0[0]
 #define vmla_lane_u32 vmla_lane_s32
 
-float32x2_t vmla_lane_f32(float32x2_t a, float32x2_t b, float32x2_t v, __constrange(0,1) int l); // VMLA.F32 d0, d0, d0[0]
+_NEON2SSESTORAGE float32x2_t vmla_lane_f32(float32x2_t a, float32x2_t b, float32x2_t v, __constrange(0,1) int l); // VMLA.F32 d0, d0, d0[0]
 _NEON2SSE_INLINE float32x2_t vmla_lane_f32(float32x2_t a, float32x2_t b, float32x2_t v, __constrange(0,1) int l)
 {
     float32_t vlane;
@@ -13329,7 +13331,7 @@ _NEON2SSE_INLINE float32x2_t vmla_lane_f32(float32x2_t a, float32x2_t b, float32
     return vmla_f32(a,b,c);
 }
 
-int16x8_t vmlaq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v, __constrange(0,3) int l); // VMLA.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE int16x8_t vmlaq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v, __constrange(0,3) int l); // VMLA.I16 q0, q0, d0[0]
 _NEON2SSE_INLINE int16x8_t vmlaq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v, __constrange(0,3) int l) // VMLA.I16 q0, q0, d0[0]
 {
     int16_t vlane;
@@ -13339,7 +13341,7 @@ _NEON2SSE_INLINE int16x8_t vmlaq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v,
     return vmlaq_s16(a,b,c);
 }
 
-int32x4_t vmlaq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v, __constrange(0,1) int l); // VMLA.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlaq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v, __constrange(0,1) int l); // VMLA.I32 q0, q0, d0[0]
 _NEON2SSE_INLINE int32x4_t vmlaq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v, __constrange(0,1) int l) // VMLA.I32 q0, q0, d0[0]
 {
     int32_t vlane;
@@ -13349,13 +13351,13 @@ _NEON2SSE_INLINE int32x4_t vmlaq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v,
     return vmlaq_s32(a,b,c);
 }
 
-uint16x8_t vmlaq_lane_u16(uint16x8_t a, uint16x8_t b, uint16x4_t v, __constrange(0,3) int l); // VMLA.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE uint16x8_t vmlaq_lane_u16(uint16x8_t a, uint16x8_t b, uint16x4_t v, __constrange(0,3) int l); // VMLA.I16 q0, q0, d0[0]
 #define vmlaq_lane_u16 vmlaq_lane_s16
 
-uint32x4_t vmlaq_lane_u32(uint32x4_t a, uint32x4_t b, uint32x2_t v, __constrange(0,1) int l); // VMLA.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlaq_lane_u32(uint32x4_t a, uint32x4_t b, uint32x2_t v, __constrange(0,1) int l); // VMLA.I32 q0, q0, d0[0]
 #define vmlaq_lane_u32 vmlaq_lane_s32
 
-float32x4_t vmlaq_lane_f32(float32x4_t a, float32x4_t b, float32x2_t v, __constrange(0,1) int l); // VMLA.F32 q0, q0, d0[0]
+_NEON2SSESTORAGE float32x4_t vmlaq_lane_f32(float32x4_t a, float32x4_t b, float32x2_t v, __constrange(0,1) int l); // VMLA.F32 q0, q0, d0[0]
 _NEON2SSE_INLINE float32x4_t vmlaq_lane_f32(float32x4_t a, float32x4_t b, float32x2_t v, __constrange(0,1) int l) // VMLA.F32 q0, q0, d0[0]
 {
     float32_t vlane;
@@ -13367,7 +13369,7 @@ _NEON2SSE_INLINE float32x4_t vmlaq_lane_f32(float32x4_t a, float32x4_t b, float3
 
 //***************** Vector widening multiply accumulate by scalar **********************
 //***************************************************************************************
-int32x4_t vmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLAL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLAL.S16 q0, d0, d0[0]
 _NEON2SSE_INLINE int32x4_t vmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l) // VMLAL.S16 q0, d0, d0[0]
 {
     int16_t vlane;
@@ -13377,7 +13379,7 @@ _NEON2SSE_INLINE int32x4_t vmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v,
     return vmlal_s16(a, b, c);
 }
 
-int64x2_t vmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLAL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLAL.S32 q0, d0, d0[0]
 _NEON2SSE_INLINE int64x2_t vmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l) // VMLAL.S32 q0, d0, d0[0]
 {
     int32_t vlane;
@@ -13387,7 +13389,7 @@ _NEON2SSE_INLINE int64x2_t vmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v,
     return vmlal_s32(a, b, c);
 }
 
-uint32x4_t vmlal_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLAL.s16 q0, d0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlal_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLAL.s16 q0, d0, d0[0]
 _NEON2SSE_INLINE uint32x4_t vmlal_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l) // VMLAL.s16 q0, d0, d0[0]
 {
     uint16_t vlane;
@@ -13397,7 +13399,7 @@ _NEON2SSE_INLINE uint32x4_t vmlal_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_
     return vmlal_u16(a, b, c);
 }
 
-uint64x2_t vmlal_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLAL.U32 q0, d0, d0[0]
+_NEON2SSESTORAGE uint64x2_t vmlal_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLAL.U32 q0, d0, d0[0]
 _NEON2SSE_INLINE uint64x2_t vmlal_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l) // VMLAL.U32 q0, d0, d0[0]
 {
     uint32_t vlane;
@@ -13409,7 +13411,7 @@ _NEON2SSE_INLINE uint64x2_t vmlal_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_
 
 // ******** Vector widening saturating doubling multiply accumulate by scalar *******************************
 // ************************************************************************************************
-int32x4_t vqdmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VQDMLAL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VQDMLAL.S16 q0, d0, d0[0]
 _NEON2SSE_INLINE int32x4_t vqdmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l)
 {
     int16_t vlane;
@@ -13419,7 +13421,7 @@ _NEON2SSE_INLINE int32x4_t vqdmlal_lane_s16(int32x4_t a, int16x4_t b, int16x4_t
     return vqdmlal_s16(a, b, c);
 }
 
-int64x2_t vqdmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VQDMLAL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VQDMLAL.S32 q0, d0, d0[0]
 _NEON2SSE_INLINE int64x2_t vqdmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l)
 {
     int32_t vlane;
@@ -13431,7 +13433,7 @@ _NEON2SSE_INLINE int64x2_t vqdmlal_lane_s32(int64x2_t a, int32x2_t b, int32x2_t
 
 // ****** Vector multiply subtract by scalar *****************
 // *************************************************************
-int16x4_t vmls_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLS.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE int16x4_t vmls_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLS.I16 d0, d0, d0[0]
 _NEON2SSE_INLINE int16x4_t vmls_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l) // VMLS.I16 d0, d0, d0[0]
 {
     int16_t vlane;
@@ -13441,7 +13443,7 @@ _NEON2SSE_INLINE int16x4_t vmls_lane_s16(int16x4_t a, int16x4_t b, int16x4_t v,
     return vmls_s16(a, b, c);
 }
 
-int32x2_t vmls_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLS.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE int32x2_t vmls_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLS.I32 d0, d0, d0[0]
 _NEON2SSE_INLINE int32x2_t vmls_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l) // VMLS.I32 d0, d0, d0[0]
 {
     int32_t vlane;
@@ -13451,7 +13453,7 @@ _NEON2SSE_INLINE int32x2_t vmls_lane_s32(int32x2_t a, int32x2_t b, int32x2_t v,
     return vmls_s32(a, b, c);
 }
 
-uint16x4_t vmls_lane_u16(uint16x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLS.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE uint16x4_t vmls_lane_u16(uint16x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLS.I16 d0, d0, d0[0]
 _NEON2SSE_INLINE uint16x4_t vmls_lane_u16(uint16x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l) // VMLS.I16 d0, d0, d0[0]
 {
     uint16_t vlane;
@@ -13461,7 +13463,7 @@ _NEON2SSE_INLINE uint16x4_t vmls_lane_u16(uint16x4_t a, uint16x4_t b, uint16x4_t
     return vmls_s16(a, b, c);
 }
 
-uint32x2_t vmls_lane_u32(uint32x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLS.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE uint32x2_t vmls_lane_u32(uint32x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLS.I32 d0, d0, d0[0]
 _NEON2SSE_INLINE uint32x2_t vmls_lane_u32(uint32x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l) // VMLS.I32 d0, d0, d0[0]
 {
     uint32_t vlane;
@@ -13471,7 +13473,7 @@ _NEON2SSE_INLINE uint32x2_t vmls_lane_u32(uint32x2_t a, uint32x2_t b, uint32x2_t
     return vmls_u32(a, b, c);
 }
 
-float32x2_t vmls_lane_f32(float32x2_t a, float32x2_t b, float32x2_t v, __constrange(0,1) int l); // VMLS.F32 d0, d0, d0[0]
+_NEON2SSESTORAGE float32x2_t vmls_lane_f32(float32x2_t a, float32x2_t b, float32x2_t v, __constrange(0,1) int l); // VMLS.F32 d0, d0, d0[0]
 _NEON2SSE_INLINE float32x2_t vmls_lane_f32(float32x2_t a, float32x2_t b, float32x2_t v, __constrange(0,1) int l)
 {
     float32_t vlane;
@@ -13481,7 +13483,7 @@ _NEON2SSE_INLINE float32x2_t vmls_lane_f32(float32x2_t a, float32x2_t b, float32
     return vmls_f32(a,b,c);
 }
 
-int16x8_t vmlsq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v, __constrange(0,3) int l); // VMLS.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE int16x8_t vmlsq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v, __constrange(0,3) int l); // VMLS.I16 q0, q0, d0[0]
 _NEON2SSE_INLINE int16x8_t vmlsq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v, __constrange(0,3) int l) // VMLS.I16 q0, q0, d0[0]
 {
     int16_t vlane;
@@ -13491,7 +13493,7 @@ _NEON2SSE_INLINE int16x8_t vmlsq_lane_s16(int16x8_t a, int16x8_t b, int16x4_t v,
     return vmlsq_s16(a, b,c);
 }
 
-int32x4_t vmlsq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v, __constrange(0,1) int l); // VMLS.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlsq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v, __constrange(0,1) int l); // VMLS.I32 q0, q0, d0[0]
 _NEON2SSE_INLINE int32x4_t vmlsq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v, __constrange(0,1) int l) // VMLS.I32 q0, q0, d0[0]
 {
     int32_t vlane;
@@ -13501,7 +13503,7 @@ _NEON2SSE_INLINE int32x4_t vmlsq_lane_s32(int32x4_t a, int32x4_t b, int32x2_t v,
     return vmlsq_s32(a,b,c);
 }
 
-uint16x8_t vmlsq_lane_u16(uint16x8_t a, uint16x8_t b, uint16x4_t v, __constrange(0,3) int l); // VMLA.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE uint16x8_t vmlsq_lane_u16(uint16x8_t a, uint16x8_t b, uint16x4_t v, __constrange(0,3) int l); // VMLA.I16 q0, q0, d0[0]
 _NEON2SSE_INLINE uint16x8_t vmlsq_lane_u16(uint16x8_t a, uint16x8_t b, uint16x4_t v, __constrange(0,3) int l) // VMLA.I16 q0, q0, d0[0]
 {
     uint16_t vlane;
@@ -13511,7 +13513,7 @@ _NEON2SSE_INLINE uint16x8_t vmlsq_lane_u16(uint16x8_t a, uint16x8_t b, uint16x4_
     return vmlsq_u16(a,b,c);
 }
 
-uint32x4_t vmlsq_lane_u32(uint32x4_t a, uint32x4_t b, uint32x2_t v, __constrange(0,1) int l); // VMLA.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlsq_lane_u32(uint32x4_t a, uint32x4_t b, uint32x2_t v, __constrange(0,1) int l); // VMLA.I32 q0, q0, d0[0]
 _NEON2SSE_INLINE uint32x4_t vmlsq_lane_u32(uint32x4_t a, uint32x4_t b, uint32x2_t v, __constrange(0,1) int l) // VMLA.I32 q0, q0, d0[0]
 {
     uint32_t vlane;
@@ -13521,7 +13523,7 @@ _NEON2SSE_INLINE uint32x4_t vmlsq_lane_u32(uint32x4_t a, uint32x4_t b, uint32x2_
     return vmlsq_u32(a,b,c);
 }
 
-float32x4_t vmlsq_lane_f32(float32x4_t a, float32x4_t b, float32x2_t v, __constrange(0,1) int l); // VMLA.F32 q0, q0, d0[0]
+_NEON2SSESTORAGE float32x4_t vmlsq_lane_f32(float32x4_t a, float32x4_t b, float32x2_t v, __constrange(0,1) int l); // VMLA.F32 q0, q0, d0[0]
 _NEON2SSE_INLINE float32x4_t vmlsq_lane_f32(float32x4_t a, float32x4_t b, float32x2_t v, __constrange(0,1) int l) // VMLA.F32 q0, q0, d0[0]
 {
     float32_t vlane;
@@ -13533,7 +13535,7 @@ _NEON2SSE_INLINE float32x4_t vmlsq_lane_f32(float32x4_t a, float32x4_t b, float3
 
 // **** Vector widening multiply subtract by scalar ****
 // ****************************************************
-int32x4_t vmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLAL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VMLAL.S16 q0, d0, d0[0]
 _NEON2SSE_INLINE int32x4_t vmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l) // VMLAL.S16 q0, d0, d0[0]
 {
     int16_t vlane;
@@ -13543,7 +13545,7 @@ _NEON2SSE_INLINE int32x4_t vmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v,
     return vmlsl_s16(a, b, c);
 }
 
-int64x2_t vmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLAL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VMLAL.S32 q0, d0, d0[0]
 _NEON2SSE_INLINE int64x2_t vmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l) // VMLAL.S32 q0, d0, d0[0]
 {
     int32_t vlane;
@@ -13553,7 +13555,7 @@ _NEON2SSE_INLINE int64x2_t vmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v,
     return vmlsl_s32(a, b, c);
 }
 
-uint32x4_t vmlsl_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLAL.s16 q0, d0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlsl_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l); // VMLAL.s16 q0, d0, d0[0]
 _NEON2SSE_INLINE uint32x4_t vmlsl_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_t v, __constrange(0,3) int l) // VMLAL.s16 q0, d0, d0[0]
 {
     uint16_t vlane;
@@ -13563,7 +13565,7 @@ _NEON2SSE_INLINE uint32x4_t vmlsl_lane_u16(uint32x4_t a, uint16x4_t b, uint16x4_
     return vmlsl_s16(a, b, c);
 }
 
-uint64x2_t vmlsl_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLAL.U32 q0, d0, d0[0]
+_NEON2SSESTORAGE uint64x2_t vmlsl_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l); // VMLAL.U32 q0, d0, d0[0]
 _NEON2SSE_INLINE uint64x2_t vmlsl_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_t v, __constrange(0,1) int l) // VMLAL.U32 q0, d0, d0[0]
 {
     uint32_t vlane;
@@ -13575,7 +13577,7 @@ _NEON2SSE_INLINE uint64x2_t vmlsl_lane_u32(uint64x2_t a, uint32x2_t b, uint32x2_
 
 //********* Vector widening saturating doubling multiply subtract by scalar **************************
 //******************************************************************************************************
-int32x4_t vqdmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VQDMLSL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l); // VQDMLSL.S16 q0, d0, d0[0]
 _NEON2SSE_INLINE int32x4_t vqdmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t v, __constrange(0,3) int l)
 {
     int16_t vlane;
@@ -13585,7 +13587,7 @@ _NEON2SSE_INLINE int32x4_t vqdmlsl_lane_s16(int32x4_t a, int16x4_t b, int16x4_t
     return vqdmlsl_s16(a, b, c);
 }
 
-int64x2_t vqdmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VQDMLSL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l); // VQDMLSL.S32 q0, d0, d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmlsl_lane_s32(int64x2_t a, int32x2_t b, int32x2_t v, __constrange(0,1) int l), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int32_t vlane;
@@ -13595,7 +13597,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmlsl_lane_s32(int64x
     return vqdmlsl_s32(a, b, c);
 }
 //********** Vector multiply with scalar *****************************
-int16x4_t vmul_n_s16(int16x4_t a, int16_t b); // VMUL.I16 d0,d0,d0[0]
+_NEON2SSESTORAGE int16x4_t vmul_n_s16(int16x4_t a, int16_t b); // VMUL.I16 d0,d0,d0[0]
 _NEON2SSE_INLINE int16x4_t vmul_n_s16(int16x4_t a, int16_t b) // VMUL.I16 d0,d0,d0[0]
 {
     int16x4_t b16x4;
@@ -13603,7 +13605,7 @@ _NEON2SSE_INLINE int16x4_t vmul_n_s16(int16x4_t a, int16_t b) // VMUL.I16 d0,d0,
     return vmul_s16(a, b16x4);
 }
 
-int32x2_t vmul_n_s32(int32x2_t a, int32_t b); // VMUL.I32 d0,d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vmul_n_s32(int32x2_t a, int32_t b); // VMUL.I32 d0,d0,d0[0]
 _NEON2SSE_INLINE int32x2_t vmul_n_s32(int32x2_t a, int32_t b) // VMUL.I32 d0,d0,d0[0]
 {
     //serial solution looks faster
@@ -13612,7 +13614,7 @@ _NEON2SSE_INLINE int32x2_t vmul_n_s32(int32x2_t a, int32_t b) // VMUL.I32 d0,d0,
     return vmul_s32(a, b32x2);
 }
 
-float32x2_t vmul_n_f32(float32x2_t a, float32_t b); // VMUL.F32 d0,d0,d0[0]
+_NEON2SSESTORAGE float32x2_t vmul_n_f32(float32x2_t a, float32_t b); // VMUL.F32 d0,d0,d0[0]
 _NEON2SSE_INLINE float32x2_t vmul_n_f32(float32x2_t a, float32_t b) // VMUL.F32 d0,d0,d0[0]
 {
     float32x2_t b32x2;
@@ -13620,7 +13622,7 @@ _NEON2SSE_INLINE float32x2_t vmul_n_f32(float32x2_t a, float32_t b) // VMUL.F32
     return vmul_f32(a, b32x2);
 }
 
-uint16x4_t vmul_n_u16(uint16x4_t a, uint16_t b); // VMUL.I16 d0,d0,d0[0]
+_NEON2SSESTORAGE uint16x4_t vmul_n_u16(uint16x4_t a, uint16_t b); // VMUL.I16 d0,d0,d0[0]
 _NEON2SSE_INLINE uint16x4_t vmul_n_u16(uint16x4_t a, uint16_t b) // VMUL.I16 d0,d0,d0[0]
 {
     uint16x4_t b16x4;
@@ -13628,7 +13630,7 @@ _NEON2SSE_INLINE uint16x4_t vmul_n_u16(uint16x4_t a, uint16_t b) // VMUL.I16 d0,
     return vmul_s16(a, b16x4);
 }
 
-uint32x2_t vmul_n_u32(uint32x2_t a, uint32_t b); // VMUL.I32 d0,d0,d0[0]
+_NEON2SSESTORAGE uint32x2_t vmul_n_u32(uint32x2_t a, uint32_t b); // VMUL.I32 d0,d0,d0[0]
 _NEON2SSE_INLINE uint32x2_t vmul_n_u32(uint32x2_t a, uint32_t b) // VMUL.I32 d0,d0,d0[0]
 {
     //serial solution looks faster
@@ -13637,7 +13639,7 @@ _NEON2SSE_INLINE uint32x2_t vmul_n_u32(uint32x2_t a, uint32_t b) // VMUL.I32 d0,
     return vmul_u32(a, b32x2);
 }
 
-int16x8_t vmulq_n_s16(int16x8_t a, int16_t b); // VMUL.I16 q0,q0,d0[0]
+_NEON2SSESTORAGE int16x8_t vmulq_n_s16(int16x8_t a, int16_t b); // VMUL.I16 q0,q0,d0[0]
 _NEON2SSE_INLINE int16x8_t vmulq_n_s16(int16x8_t a, int16_t b) // VMUL.I16 q0,q0,d0[0]
 {
     int16x8_t b16x8;
@@ -13645,7 +13647,7 @@ _NEON2SSE_INLINE int16x8_t vmulq_n_s16(int16x8_t a, int16_t b) // VMUL.I16 q0,q0
     return vmulq_s16(a, b16x8);
 }
 
-int32x4_t vmulq_n_s32(int32x4_t a, int32_t b); // VMUL.I32 q0,q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vmulq_n_s32(int32x4_t a, int32_t b); // VMUL.I32 q0,q0,d0[0]
 _NEON2SSE_INLINE int32x4_t vmulq_n_s32(int32x4_t a, int32_t b) // VMUL.I32 q0,q0,d0[0]
 {
     int32x4_t b32x4;
@@ -13653,7 +13655,7 @@ _NEON2SSE_INLINE int32x4_t vmulq_n_s32(int32x4_t a, int32_t b) // VMUL.I32 q0,q0
     return vmulq_s32(a, b32x4);
 }
 
-float32x4_t vmulq_n_f32(float32x4_t a, float32_t b); // VMUL.F32 q0,q0,d0[0]
+_NEON2SSESTORAGE float32x4_t vmulq_n_f32(float32x4_t a, float32_t b); // VMUL.F32 q0,q0,d0[0]
 _NEON2SSE_INLINE float32x4_t vmulq_n_f32(float32x4_t a, float32_t b) // VMUL.F32 q0,q0,d0[0]
 {
     float32x4_t b32x4;
@@ -13661,7 +13663,7 @@ _NEON2SSE_INLINE float32x4_t vmulq_n_f32(float32x4_t a, float32_t b) // VMUL.F32
     return vmulq_f32(a, b32x4);
 }
 
-uint16x8_t vmulq_n_u16(uint16x8_t a, uint16_t b); // VMUL.I16 q0,q0,d0[0]
+_NEON2SSESTORAGE uint16x8_t vmulq_n_u16(uint16x8_t a, uint16_t b); // VMUL.I16 q0,q0,d0[0]
 _NEON2SSE_INLINE uint16x8_t vmulq_n_u16(uint16x8_t a, uint16_t b) // VMUL.I16 q0,q0,d0[0]
 {
     uint16x8_t b16x8;
@@ -13669,7 +13671,7 @@ _NEON2SSE_INLINE uint16x8_t vmulq_n_u16(uint16x8_t a, uint16_t b) // VMUL.I16 q0
     return vmulq_s16(a, b16x8);
 }
 
-uint32x4_t vmulq_n_u32(uint32x4_t a, uint32_t b); // VMUL.I32 q0,q0,d0[0]
+_NEON2SSESTORAGE uint32x4_t vmulq_n_u32(uint32x4_t a, uint32_t b); // VMUL.I32 q0,q0,d0[0]
 _NEON2SSE_INLINE uint32x4_t vmulq_n_u32(uint32x4_t a, uint32_t b) // VMUL.I32 q0,q0,d0[0]
 {
     uint32x4_t b32x4;
@@ -13678,7 +13680,7 @@ _NEON2SSE_INLINE uint32x4_t vmulq_n_u32(uint32x4_t a, uint32_t b) // VMUL.I32 q0
 }
 
 //********** Vector multiply lane *****************************
-int16x4_t vmul_lane_s16 (int16x4_t a, int16x4_t b, __constrange(0,3) int c);
+_NEON2SSESTORAGE int16x4_t vmul_lane_s16 (int16x4_t a, int16x4_t b, __constrange(0,3) int c);
 _NEON2SSE_INLINE int16x4_t vmul_lane_s16 (int16x4_t a, int16x4_t b, __constrange(0,3) int c)
 {
     int16x4_t b16x4;
@@ -13688,7 +13690,7 @@ _NEON2SSE_INLINE int16x4_t vmul_lane_s16 (int16x4_t a, int16x4_t b, __constrange
     return vmul_s16(a, b16x4);
 }
 
-int32x2_t vmul_lane_s32 (int32x2_t a, int32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE int32x2_t vmul_lane_s32 (int32x2_t a, int32x2_t b, __constrange(0,1) int c);
 _NEON2SSE_INLINE int32x2_t vmul_lane_s32 (int32x2_t a, int32x2_t b, __constrange(0,1) int c)
 {
     int32x2_t b32x2;
@@ -13698,7 +13700,7 @@ _NEON2SSE_INLINE int32x2_t vmul_lane_s32 (int32x2_t a, int32x2_t b, __constrange
     return vmul_s32(a, b32x2);
 }
 
-float32x2_t vmul_lane_f32 (float32x2_t a, float32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE float32x2_t vmul_lane_f32 (float32x2_t a, float32x2_t b, __constrange(0,1) int c);
 _NEON2SSE_INLINE float32x2_t vmul_lane_f32 (float32x2_t a, float32x2_t b, __constrange(0,1) int c)
 {
     float32x2_t b32x2;
@@ -13708,13 +13710,13 @@ _NEON2SSE_INLINE float32x2_t vmul_lane_f32 (float32x2_t a, float32x2_t b, __cons
     return vmul_f32(a, b32x2);
 }
 
-uint16x4_t vmul_lane_u16 (uint16x4_t a, uint16x4_t b, __constrange(0,3) int c);
+_NEON2SSESTORAGE uint16x4_t vmul_lane_u16 (uint16x4_t a, uint16x4_t b, __constrange(0,3) int c);
 #define vmul_lane_u16 vmul_lane_s16
 
-uint32x2_t vmul_lane_u32 (uint32x2_t a, uint32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE uint32x2_t vmul_lane_u32 (uint32x2_t a, uint32x2_t b, __constrange(0,1) int c);
 #define vmul_lane_u32 vmul_lane_s32
 
-int16x8_t vmulq_lane_s16(int16x8_t a, int16x4_t b, __constrange(0,3) int c);
+_NEON2SSESTORAGE int16x8_t vmulq_lane_s16(int16x8_t a, int16x4_t b, __constrange(0,3) int c);
 _NEON2SSE_INLINE int16x8_t vmulq_lane_s16 (int16x8_t a, int16x4_t b, __constrange(0,3) int c)
 {
     int16x8_t b16x8;
@@ -13724,7 +13726,7 @@ _NEON2SSE_INLINE int16x8_t vmulq_lane_s16 (int16x8_t a, int16x4_t b, __constrang
     return vmulq_s16(a, b16x8);
 }
 
-int32x4_t vmulq_lane_s32 (int32x4_t a, int32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE int32x4_t vmulq_lane_s32 (int32x4_t a, int32x2_t b, __constrange(0,1) int c);
 _NEON2SSE_INLINE int32x4_t vmulq_lane_s32 (int32x4_t a, int32x2_t b, __constrange(0,1) int c)
 {
     int32x4_t b32x4;
@@ -13734,7 +13736,7 @@ _NEON2SSE_INLINE int32x4_t vmulq_lane_s32 (int32x4_t a, int32x2_t b, __constrang
     return vmulq_s32(a, b32x4);
 }
 
-float32x4_t vmulq_lane_f32 (float32x4_t a, float32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE float32x4_t vmulq_lane_f32 (float32x4_t a, float32x2_t b, __constrange(0,1) int c);
 _NEON2SSE_INLINE float32x4_t vmulq_lane_f32 (float32x4_t a, float32x2_t b, __constrange(0,1) int c)
 {
     float32x4_t b32x4;
@@ -13744,14 +13746,14 @@ _NEON2SSE_INLINE float32x4_t vmulq_lane_f32 (float32x4_t a, float32x2_t b, __con
     return vmulq_f32(a, b32x4);
 }
 
-uint16x8_t vmulq_lane_u16 (uint16x8_t a, uint16x4_t b, __constrange(0,3) int c);
+_NEON2SSESTORAGE uint16x8_t vmulq_lane_u16 (uint16x8_t a, uint16x4_t b, __constrange(0,3) int c);
 #define vmulq_lane_u16 vmulq_lane_s16
 
-uint32x4_t vmulq_lane_u32 (uint32x4_t a, uint32x2_t b, __constrange(0,1) int c);
+_NEON2SSESTORAGE uint32x4_t vmulq_lane_u32 (uint32x4_t a, uint32x2_t b, __constrange(0,1) int c);
 #define vmulq_lane_u32 vmulq_lane_s32
 
 //**** Vector long multiply with scalar ************
-int32x4_t vmull_n_s16(int16x4_t vec1, int16_t val2); // VMULL.S16 q0,d0,d0[0]
+_NEON2SSESTORAGE int32x4_t vmull_n_s16(int16x4_t vec1, int16_t val2); // VMULL.S16 q0,d0,d0[0]
 _NEON2SSE_INLINE int32x4_t vmull_n_s16(int16x4_t vec1, int16_t val2) // VMULL.S16 q0,d0,d0[0]
 {
     int16x4_t b16x4;
@@ -13759,7 +13761,7 @@ _NEON2SSE_INLINE int32x4_t vmull_n_s16(int16x4_t vec1, int16_t val2) // VMULL.S1
     return vmull_s16(vec1, b16x4);
 }
 
-int64x2_t vmull_n_s32(int32x2_t vec1, int32_t val2); // VMULL.S32 q0,d0,d0[0]
+_NEON2SSESTORAGE int64x2_t vmull_n_s32(int32x2_t vec1, int32_t val2); // VMULL.S32 q0,d0,d0[0]
 _NEON2SSE_INLINE int64x2_t vmull_n_s32(int32x2_t vec1, int32_t val2) // VMULL.S32 q0,d0,d0[0]
 {
     int32x2_t b32x2;
@@ -13767,7 +13769,7 @@ _NEON2SSE_INLINE int64x2_t vmull_n_s32(int32x2_t vec1, int32_t val2) // VMULL.S3
     return vmull_s32(vec1, b32x2);
 }
 
-uint32x4_t vmull_n_u16(uint16x4_t vec1, uint16_t val2); // VMULL.s16 q0,d0,d0[0]
+_NEON2SSESTORAGE uint32x4_t vmull_n_u16(uint16x4_t vec1, uint16_t val2); // VMULL.s16 q0,d0,d0[0]
 _NEON2SSE_INLINE uint32x4_t vmull_n_u16(uint16x4_t vec1, uint16_t val2) // VMULL.s16 q0,d0,d0[0]
 {
     uint16x4_t b16x4;
@@ -13775,7 +13777,7 @@ _NEON2SSE_INLINE uint32x4_t vmull_n_u16(uint16x4_t vec1, uint16_t val2) // VMULL
     return vmull_s16(vec1, b16x4);
 }
 
-uint64x2_t vmull_n_u32(uint32x2_t vec1, uint32_t val2); // VMULL.U32 q0,d0,d0[0]
+_NEON2SSESTORAGE uint64x2_t vmull_n_u32(uint32x2_t vec1, uint32_t val2); // VMULL.U32 q0,d0,d0[0]
 _NEON2SSE_INLINE uint64x2_t vmull_n_u32(uint32x2_t vec1, uint32_t val2) // VMULL.U32 q0,d0,d0[0]
 {
     uint32x2_t b32x2;
@@ -13784,7 +13786,7 @@ _NEON2SSE_INLINE uint64x2_t vmull_n_u32(uint32x2_t vec1, uint32_t val2) // VMULL
 }
 
 //**** Vector long multiply by scalar ****
-int32x4_t vmull_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VMULL.S16 q0,d0,d0[0]
+_NEON2SSESTORAGE int32x4_t vmull_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VMULL.S16 q0,d0,d0[0]
 _NEON2SSE_INLINE int32x4_t vmull_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3) // VMULL.S16 q0,d0,d0[0]
 {
     int16_t vlane;
@@ -13794,7 +13796,7 @@ _NEON2SSE_INLINE int32x4_t vmull_lane_s16(int16x4_t vec1, int16x4_t val2, __cons
     return vmull_s16(vec1, b);
 }
 
-int64x2_t vmull_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VMULL.S32 q0,d0,d0[0]
+_NEON2SSESTORAGE int64x2_t vmull_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VMULL.S32 q0,d0,d0[0]
 _NEON2SSE_INLINE int64x2_t vmull_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3) // VMULL.S32 q0,d0,d0[0]
 {
     int32_t vlane;
@@ -13804,7 +13806,7 @@ _NEON2SSE_INLINE int64x2_t vmull_lane_s32(int32x2_t vec1, int32x2_t val2, __cons
     return vmull_s32(vec1, b);
 }
 
-uint32x4_t vmull_lane_u16(uint16x4_t vec1, uint16x4_t val2, __constrange(0, 3) int val3); // VMULL.s16 q0,d0,d0[0]
+_NEON2SSESTORAGE uint32x4_t vmull_lane_u16(uint16x4_t vec1, uint16x4_t val2, __constrange(0, 3) int val3); // VMULL.s16 q0,d0,d0[0]
 _NEON2SSE_INLINE uint32x4_t vmull_lane_u16(uint16x4_t vec1, uint16x4_t val2, __constrange(0, 3) int val3) // VMULL.s16 q0,d0,d0[0]
 {
     uint16_t vlane;
@@ -13814,7 +13816,7 @@ _NEON2SSE_INLINE uint32x4_t vmull_lane_u16(uint16x4_t vec1, uint16x4_t val2, __c
     return vmull_s16(vec1, b);
 }
 
-uint64x2_t vmull_lane_u32(uint32x2_t vec1, uint32x2_t val2, __constrange(0, 1) int val3); // VMULL.U32 q0,d0,d0[0]
+_NEON2SSESTORAGE uint64x2_t vmull_lane_u32(uint32x2_t vec1, uint32x2_t val2, __constrange(0, 1) int val3); // VMULL.U32 q0,d0,d0[0]
 _NEON2SSE_INLINE uint64x2_t vmull_lane_u32(uint32x2_t vec1, uint32x2_t val2, __constrange(0, 1) int val3) // VMULL.U32 q0,d0,d0[0]
 {
     uint32_t vlane;
@@ -13825,7 +13827,7 @@ _NEON2SSE_INLINE uint64x2_t vmull_lane_u32(uint32x2_t vec1, uint32x2_t val2, __c
 }
 
 //********* Vector saturating doubling long multiply with scalar  *******************
-int32x4_t vqdmull_n_s16(int16x4_t vec1, int16_t val2); // VQDMULL.S16 q0,d0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmull_n_s16(int16x4_t vec1, int16_t val2); // VQDMULL.S16 q0,d0,d0[0]
 _NEON2SSE_INLINE int32x4_t vqdmull_n_s16(int16x4_t vec1, int16_t val2)
 {
     //the serial soulution may be faster due to saturation
@@ -13834,7 +13836,7 @@ _NEON2SSE_INLINE int32x4_t vqdmull_n_s16(int16x4_t vec1, int16_t val2)
     return vqdmull_s16(vec1, b);
 }
 
-int64x2_t vqdmull_n_s32(int32x2_t vec1, int32_t val2); // VQDMULL.S32 q0,d0,d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmull_n_s32(int32x2_t vec1, int32_t val2); // VQDMULL.S32 q0,d0,d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmull_n_s32(int32x2_t vec1, int32_t val2), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int32x2_t b;
@@ -13843,7 +13845,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmull_n_s32(int32x2_t
 }
 
 //************* Vector saturating doubling long multiply by scalar ***********************************************
-int32x4_t vqdmull_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQDMULL.S16 q0,d0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmull_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQDMULL.S16 q0,d0,d0[0]
 _NEON2SSE_INLINE int32x4_t vqdmull_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3)
 {
     int16_t c;
@@ -13854,7 +13856,7 @@ _NEON2SSE_INLINE int32x4_t vqdmull_lane_s16(int16x4_t vec1, int16x4_t val2, __co
 }
 
 
-int64x2_t vqdmull_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); //  VQDMULL.S32 q0,d0,d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmull_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); //  VQDMULL.S32 q0,d0,d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmull_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int32_t c;
@@ -13865,21 +13867,21 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmull_lane_s32(int32x
 }
 
 // *****Vector saturating doubling multiply high with scalar *****
-int16x4_t vqdmulh_n_s16(int16x4_t vec1,  int16_t val2); //  VQDMULH.S16 d0,d0,d0[0]
+_NEON2SSESTORAGE int16x4_t vqdmulh_n_s16(int16x4_t vec1,  int16_t val2); //  VQDMULH.S16 d0,d0,d0[0]
 _NEON2SSE_INLINE int16x4_t vqdmulh_n_s16(int16x4_t vec1,  int16_t val2)
 {
     int16x4_t res64;
     return64(vqdmulhq_n_s16(_pM128i(vec1), val2));
 }
 
-int32x2_t vqdmulh_n_s32(int32x2_t vec1,  int32_t val2); //  VQDMULH.S32 d0,d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vqdmulh_n_s32(int32x2_t vec1,  int32_t val2); //  VQDMULH.S32 d0,d0,d0[0]
 _NEON2SSE_INLINE int32x2_t vqdmulh_n_s32(int32x2_t vec1,  int32_t val2)
 {
     int32x2_t res64;
     return64(vqdmulhq_n_s32(_pM128i(vec1), val2));
 }
 
-int16x8_t vqdmulhq_n_s16(int16x8_t vec1, int16_t val2); //  VQDMULH.S16 q0,q0,d0[0]
+_NEON2SSESTORAGE int16x8_t vqdmulhq_n_s16(int16x8_t vec1, int16_t val2); //  VQDMULH.S16 q0,q0,d0[0]
 _NEON2SSE_INLINE int16x8_t vqdmulhq_n_s16(int16x8_t vec1, int16_t val2) //  VQDMULH.S16 q0,q0,d0[0]
 {
     //solution may be not optimal
@@ -13888,7 +13890,7 @@ _NEON2SSE_INLINE int16x8_t vqdmulhq_n_s16(int16x8_t vec1, int16_t val2) //  VQDM
     return vqdmulhq_s16(vec1, scalar);
 }
 
-int32x4_t vqdmulhq_n_s32(int32x4_t vec1, int32_t val2); //  VQDMULH.S32 q0,q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmulhq_n_s32(int32x4_t vec1, int32_t val2); //  VQDMULH.S32 q0,q0,d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqdmulhq_n_s32(int32x4_t vec1, int32_t val2), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     int32x4_t scalar;
@@ -13897,7 +13899,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqdmulhq_n_s32(int32x4_
 }
 
 //***** Vector saturating doubling multiply high by scalar ****************
-int16x4_t vqdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); //  VQDMULH.S16 d0,d0,d0[0]
+_NEON2SSESTORAGE int16x4_t vqdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); //  VQDMULH.S16 d0,d0,d0[0]
 _NEON2SSE_INLINE int16x4_t vqdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3) //  VQDMULH.S16 d0,d0,d0[0]
 {
     //solution may be not optimal
@@ -13908,7 +13910,7 @@ _NEON2SSE_INLINE int16x4_t vqdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __co
     return vqdmulh_s16(vec1, scalar);
 }
 
-int32x2_t vqdmulh_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); //  VQDMULH.S32 d0,d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vqdmulh_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); //  VQDMULH.S32 d0,d0,d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqdmulh_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     int32_t vlane;
@@ -13918,7 +13920,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqdmulh_lane_s32(int32x
     return vqdmulh_s32(vec1, scalar);
 }
 
-int16x8_t vqdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __constrange(0, 3) int val3); //  VQDMULH.S16 q0,q0,d0[0]
+_NEON2SSESTORAGE int16x8_t vqdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __constrange(0, 3) int val3); //  VQDMULH.S16 q0,q0,d0[0]
 _NEON2SSE_INLINE int16x8_t vqdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __constrange(0, 3) int val3) //  VQDMULH.S16 q0,q0,d0[0]
 {
     //solution may be not optimal
@@ -13929,7 +13931,7 @@ _NEON2SSE_INLINE int16x8_t vqdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __c
     return vqdmulhq_s16(vec1, scalar);
 }
 
-int32x4_t vqdmulhq_lane_s32(int32x4_t vec1, int32x2_t val2, __constrange(0, 1) int val3); //  VQDMULH.S32 q0,q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmulhq_lane_s32(int32x4_t vec1, int32x2_t val2, __constrange(0, 1) int val3); //  VQDMULH.S32 q0,q0,d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqdmulhq_lane_s32(int32x4_t vec1, int32x2_t val2, __constrange(0, 1) int val3), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     //solution may be not optimal
@@ -13941,7 +13943,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqdmulhq_lane_s32(int32
 }
 
 //******** Vector saturating rounding doubling multiply high with scalar ***
-int16x4_t vqrdmulh_n_s16(int16x4_t vec1, int16_t val2); // VQRDMULH.S16 d0,d0,d0[0]
+_NEON2SSESTORAGE int16x4_t vqrdmulh_n_s16(int16x4_t vec1, int16_t val2); // VQRDMULH.S16 d0,d0,d0[0]
 _NEON2SSE_INLINE int16x4_t vqrdmulh_n_s16(int16x4_t vec1, int16_t val2) // VQRDMULH.S16 d0,d0,d0[0]
 {
     //solution may be not optimal
@@ -13950,7 +13952,7 @@ _NEON2SSE_INLINE int16x4_t vqrdmulh_n_s16(int16x4_t vec1, int16_t val2) // VQRDM
     return vqrdmulh_s16(vec1, scalar);
 }
 
-int32x2_t vqrdmulh_n_s32(int32x2_t vec1, int32_t val2); // VQRDMULH.S32 d0,d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vqrdmulh_n_s32(int32x2_t vec1, int32_t val2); // VQRDMULH.S32 d0,d0,d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqrdmulh_n_s32(int32x2_t vec1, int32_t val2), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     int32x2_t scalar;
@@ -13958,7 +13960,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqrdmulh_n_s32(int32x2_
     return vqrdmulh_s32(vec1, scalar);
 }
 
-int16x8_t vqrdmulhq_n_s16(int16x8_t vec1, int16_t val2); // VQRDMULH.S16 q0,q0,d0[0]
+_NEON2SSESTORAGE int16x8_t vqrdmulhq_n_s16(int16x8_t vec1, int16_t val2); // VQRDMULH.S16 q0,q0,d0[0]
 _NEON2SSE_INLINE int16x8_t vqrdmulhq_n_s16(int16x8_t vec1, int16_t val2) // VQRDMULH.S16 q0,q0,d0[0]
 {
     //solution may be not optimal
@@ -13967,7 +13969,7 @@ _NEON2SSE_INLINE int16x8_t vqrdmulhq_n_s16(int16x8_t vec1, int16_t val2) // VQRD
     return vqrdmulhq_s16(vec1, scalar);
 }
 
-int32x4_t vqrdmulhq_n_s32(int32x4_t vec1, int32_t val2); // VQRDMULH.S32 q0,q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqrdmulhq_n_s32(int32x4_t vec1, int32_t val2); // VQRDMULH.S32 q0,q0,d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqrdmulhq_n_s32(int32x4_t vec1, int32_t val2), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     int32x4_t scalar;
@@ -13976,7 +13978,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqrdmulhq_n_s32(int32x4
 }
 
 //********* Vector rounding saturating doubling multiply high by scalar  ****
-int16x4_t vqrdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQRDMULH.S16 d0,d0,d0[0]
+_NEON2SSESTORAGE int16x4_t vqrdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQRDMULH.S16 d0,d0,d0[0]
 _NEON2SSE_INLINE int16x4_t vqrdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __constrange(0, 3) int val3) // VQRDMULH.S16 d0,d0,d0[0]
 {
     //solution may be not optimal
@@ -13987,7 +13989,7 @@ _NEON2SSE_INLINE int16x4_t vqrdmulh_lane_s16(int16x4_t vec1, int16x4_t val2, __c
     return vqrdmulh_s16(vec1, scalar);
 }
 
-int32x2_t vqrdmulh_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQRDMULH.S32 d0,d0,d0[0]
+_NEON2SSESTORAGE int32x2_t vqrdmulh_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQRDMULH.S32 d0,d0,d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqrdmulh_lane_s32(int32x2_t vec1, int32x2_t val2, __constrange(0, 1) int val3), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     int32_t vlane;
@@ -13997,7 +13999,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vqrdmulh_lane_s32(int32
     return vqrdmulh_s32(vec1, scalar);
 }
 
-int16x8_t vqrdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQRDMULH.S16 q0,q0,d0[0]
+_NEON2SSESTORAGE int16x8_t vqrdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __constrange(0, 3) int val3); // VQRDMULH.S16 q0,q0,d0[0]
 _NEON2SSE_INLINE int16x8_t vqrdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __constrange(0, 3) int val3) // VQRDMULH.S16 q0,q0,d0[0]
 {
     //solution may be not optimal
@@ -14008,7 +14010,7 @@ _NEON2SSE_INLINE int16x8_t vqrdmulhq_lane_s16(int16x8_t vec1, int16x4_t val2, __
     return vqrdmulhq_s16(vec1, scalar);
 }
 
-int32x4_t vqrdmulhq_lane_s32(int32x4_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQRDMULH.S32 q0,q0,d0[0]
+_NEON2SSESTORAGE int32x4_t vqrdmulhq_lane_s32(int32x4_t vec1, int32x2_t val2, __constrange(0, 1) int val3); // VQRDMULH.S32 q0,q0,d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqrdmulhq_lane_s32(int32x4_t vec1, int32x2_t val2, __constrange(0, 1) int val3), _NEON2SSE_REASON_SLOW_UNEFFECTIVE)
 {
     //solution may be not optimal
@@ -14020,7 +14022,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x4_t vqrdmulhq_lane_s32(int3
 }
 
 //**************Vector multiply accumulate with scalar *******************
-int16x4_t vmla_n_s16(int16x4_t a, int16x4_t b, int16_t c); // VMLA.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE int16x4_t vmla_n_s16(int16x4_t a, int16x4_t b, int16_t c); // VMLA.I16 d0, d0, d0[0]
 _NEON2SSE_INLINE int16x4_t vmla_n_s16(int16x4_t a, int16x4_t b, int16_t c) // VMLA.I16 d0, d0, d0[0]
 {
     int16x4_t scalar;
@@ -14028,7 +14030,7 @@ _NEON2SSE_INLINE int16x4_t vmla_n_s16(int16x4_t a, int16x4_t b, int16_t c) // VM
     return vmla_s16(a, b, scalar);
 }
 
-int32x2_t vmla_n_s32(int32x2_t a, int32x2_t b, int32_t c); // VMLA.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE int32x2_t vmla_n_s32(int32x2_t a, int32x2_t b, int32_t c); // VMLA.I32 d0, d0, d0[0]
 _NEON2SSE_INLINE int32x2_t vmla_n_s32(int32x2_t a, int32x2_t b, int32_t c) // VMLA.I32 d0, d0, d0[0]
 {
     int32x2_t scalar;
@@ -14036,15 +14038,15 @@ _NEON2SSE_INLINE int32x2_t vmla_n_s32(int32x2_t a, int32x2_t b, int32_t c) // VM
     return vmla_s32(a, b, scalar);
 }
 
-uint16x4_t vmla_n_u16(uint16x4_t a,  uint16x4_t b, uint16_t c); // VMLA.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE uint16x4_t vmla_n_u16(uint16x4_t a,  uint16x4_t b, uint16_t c); // VMLA.I16 d0, d0, d0[0]
 #define vmla_n_u16 vmla_n_s16
 
 
-uint32x2_t vmla_n_u32(uint32x2_t a,  uint32x2_t b, uint32_t c); // VMLA.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE uint32x2_t vmla_n_u32(uint32x2_t a,  uint32x2_t b, uint32_t c); // VMLA.I32 d0, d0, d0[0]
 #define vmla_n_u32 vmla_n_s32
 
 
-float32x2_t vmla_n_f32(float32x2_t a, float32x2_t b, float32_t c); // VMLA.F32 d0, d0, d0[0]
+_NEON2SSESTORAGE float32x2_t vmla_n_f32(float32x2_t a, float32x2_t b, float32_t c); // VMLA.F32 d0, d0, d0[0]
 _NEON2SSE_INLINE float32x2_t vmla_n_f32(float32x2_t a, float32x2_t b, float32_t c) // VMLA.F32 d0, d0, d0[0]
 {
     float32x2_t scalar;
@@ -14052,7 +14054,7 @@ _NEON2SSE_INLINE float32x2_t vmla_n_f32(float32x2_t a, float32x2_t b, float32_t
     return vmla_f32(a, b, scalar);
 }
 
-int16x8_t vmlaq_n_s16(int16x8_t a, int16x8_t b, int16_t c); // VMLA.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE int16x8_t vmlaq_n_s16(int16x8_t a, int16x8_t b, int16_t c); // VMLA.I16 q0, q0, d0[0]
 _NEON2SSE_INLINE int16x8_t vmlaq_n_s16(int16x8_t a, int16x8_t b, int16_t c) // VMLA.I16 q0, q0, d0[0]
 {
     int16x8_t scalar;
@@ -14060,7 +14062,7 @@ _NEON2SSE_INLINE int16x8_t vmlaq_n_s16(int16x8_t a, int16x8_t b, int16_t c) // V
     return vmlaq_s16(a,b,scalar);
 }
 
-int32x4_t vmlaq_n_s32(int32x4_t a, int32x4_t b, int32_t c); // VMLA.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlaq_n_s32(int32x4_t a, int32x4_t b, int32_t c); // VMLA.I32 q0, q0, d0[0]
 _NEON2SSE_INLINE int32x4_t vmlaq_n_s32(int32x4_t a, int32x4_t b, int32_t c) // VMLA.I32 q0, q0, d0[0]
 {
     int32x4_t scalar;
@@ -14068,13 +14070,13 @@ _NEON2SSE_INLINE int32x4_t vmlaq_n_s32(int32x4_t a, int32x4_t b, int32_t c) // V
     return vmlaq_s32(a,b,scalar);
 }
 
-uint16x8_t vmlaq_n_u16(uint16x8_t a, uint16x8_t b, uint16_t c); // VMLA.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE uint16x8_t vmlaq_n_u16(uint16x8_t a, uint16x8_t b, uint16_t c); // VMLA.I16 q0, q0, d0[0]
 #define vmlaq_n_u16 vmlaq_n_s16
 
-uint32x4_t vmlaq_n_u32(uint32x4_t a, uint32x4_t b, uint32_t c); // VMLA.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlaq_n_u32(uint32x4_t a, uint32x4_t b, uint32_t c); // VMLA.I32 q0, q0, d0[0]
 #define vmlaq_n_u32 vmlaq_n_s32
 
-float32x4_t vmlaq_n_f32(float32x4_t a, float32x4_t b, float32_t c); // VMLA.F32 q0, q0, d0[0]
+_NEON2SSESTORAGE float32x4_t vmlaq_n_f32(float32x4_t a, float32x4_t b, float32_t c); // VMLA.F32 q0, q0, d0[0]
 _NEON2SSE_INLINE float32x4_t vmlaq_n_f32(float32x4_t a, float32x4_t b, float32_t c) // VMLA.F32 q0, q0, d0[0]
 {
     float32x4_t scalar;
@@ -14083,7 +14085,7 @@ _NEON2SSE_INLINE float32x4_t vmlaq_n_f32(float32x4_t a, float32x4_t b, float32_t
 }
 
 //************Vector widening multiply accumulate with scalar****************************
-int32x4_t vmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VMLAL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VMLAL.S16 q0, d0, d0[0]
 _NEON2SSE_INLINE int32x4_t vmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c) // VMLAL.S16 q0, d0, d0[0]
 {
     int16x4_t vc;
@@ -14091,7 +14093,7 @@ _NEON2SSE_INLINE int32x4_t vmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c) // V
     return vmlal_s16(a, b, vc);
 }
 
-int64x2_t vmlal_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VMLAL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vmlal_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VMLAL.S32 q0, d0, d0[0]
 _NEON2SSE_INLINE int64x2_t vmlal_n_s32(int64x2_t a, int32x2_t b, int32_t c) // VMLAL.S32 q0, d0, d0[0]
 {
     int32x2_t vc;
@@ -14099,7 +14101,7 @@ _NEON2SSE_INLINE int64x2_t vmlal_n_s32(int64x2_t a, int32x2_t b, int32_t c) // V
     return vmlal_s32(a, b, vc);
 }
 
-uint32x4_t vmlal_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c); // VMLAL.s16 q0, d0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlal_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c); // VMLAL.s16 q0, d0, d0[0]
 _NEON2SSE_INLINE uint32x4_t vmlal_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c) // VMLAL.s16 q0, d0, d0[0]
 {
     uint16x4_t vc;
@@ -14107,7 +14109,7 @@ _NEON2SSE_INLINE uint32x4_t vmlal_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c)
     return vmlal_u16(a, b, vc);
 }
 
-uint64x2_t vmlal_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c); // VMLAL.U32 q0, d0, d0[0]
+_NEON2SSESTORAGE uint64x2_t vmlal_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c); // VMLAL.U32 q0, d0, d0[0]
 _NEON2SSE_INLINE uint64x2_t vmlal_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c) // VMLAL.U32 q0, d0, d0[0]
 {
     uint32x2_t vc;
@@ -14116,7 +14118,7 @@ _NEON2SSE_INLINE uint64x2_t vmlal_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c)
 }
 
 //************ Vector widening saturating doubling multiply accumulate with scalar **************
-int32x4_t vqdmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VQDMLAL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VQDMLAL.S16 q0, d0, d0[0]
 _NEON2SSE_INLINE int32x4_t vqdmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c)
 {
     //not optimal SIMD soulution, serial may be faster
@@ -14125,7 +14127,7 @@ _NEON2SSE_INLINE int32x4_t vqdmlal_n_s16(int32x4_t a, int16x4_t b, int16_t c)
     return vqdmlal_s16(a, b, vc);
 }
 
-int64x2_t vqdmlal_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VQDMLAL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmlal_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VQDMLAL.S32 q0, d0, d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmlal_n_s32(int64x2_t a, int32x2_t b, int32_t c), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int32x2_t vc;
@@ -14134,7 +14136,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmlal_n_s32(int64x2_t
 }
 
 //******** Vector multiply subtract with scalar **************
-int16x4_t vmls_n_s16(int16x4_t a, int16x4_t b, int16_t c); // VMLS.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE int16x4_t vmls_n_s16(int16x4_t a, int16x4_t b, int16_t c); // VMLS.I16 d0, d0, d0[0]
 _NEON2SSE_INLINE int16x4_t vmls_n_s16(int16x4_t a, int16x4_t b, int16_t c) // VMLS.I16 d0, d0, d0[0]
 {
     int16x4_t vc;
@@ -14142,7 +14144,7 @@ _NEON2SSE_INLINE int16x4_t vmls_n_s16(int16x4_t a, int16x4_t b, int16_t c) // VM
     return vmls_s16(a, b, vc);
 }
 
-int32x2_t vmls_n_s32(int32x2_t a, int32x2_t b, int32_t c); // VMLS.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE int32x2_t vmls_n_s32(int32x2_t a, int32x2_t b, int32_t c); // VMLS.I32 d0, d0, d0[0]
 _NEON2SSE_INLINE int32x2_t vmls_n_s32(int32x2_t a, int32x2_t b, int32_t c) // VMLS.I32 d0, d0, d0[0]
 {
     int32x2_t vc;
@@ -14150,7 +14152,7 @@ _NEON2SSE_INLINE int32x2_t vmls_n_s32(int32x2_t a, int32x2_t b, int32_t c) // VM
     return vmls_s32(a, b, vc);
 }
 
-uint16x4_t vmls_n_u16(uint16x4_t a, uint16x4_t b, uint16_t c); // VMLS.I16 d0, d0, d0[0]
+_NEON2SSESTORAGE uint16x4_t vmls_n_u16(uint16x4_t a, uint16x4_t b, uint16_t c); // VMLS.I16 d0, d0, d0[0]
 _NEON2SSE_INLINE uint16x4_t vmls_n_u16(uint16x4_t a, uint16x4_t b, uint16_t c) // VMLS.I16 d0, d0, d0[0]
 {
     uint16x4_t vc;
@@ -14158,7 +14160,7 @@ _NEON2SSE_INLINE uint16x4_t vmls_n_u16(uint16x4_t a, uint16x4_t b, uint16_t c) /
     return vmls_s16(a, b, vc);
 }
 
-uint32x2_t vmls_n_u32(uint32x2_t a, uint32x2_t b, uint32_t c); // VMLS.I32 d0, d0, d0[0]
+_NEON2SSESTORAGE uint32x2_t vmls_n_u32(uint32x2_t a, uint32x2_t b, uint32_t c); // VMLS.I32 d0, d0, d0[0]
 _NEON2SSE_INLINE uint32x2_t vmls_n_u32(uint32x2_t a, uint32x2_t b, uint32_t c) // VMLS.I32 d0, d0, d0[0]
 {
     uint32x2_t vc;
@@ -14166,7 +14168,7 @@ _NEON2SSE_INLINE uint32x2_t vmls_n_u32(uint32x2_t a, uint32x2_t b, uint32_t c) /
     return vmls_u32(a, b, vc);
 }
 
-float32x2_t vmls_n_f32(float32x2_t a, float32x2_t b, float32_t c); // VMLS.F32 d0, d0, d0[0]
+_NEON2SSESTORAGE float32x2_t vmls_n_f32(float32x2_t a, float32x2_t b, float32_t c); // VMLS.F32 d0, d0, d0[0]
 _NEON2SSE_INLINE float32x2_t vmls_n_f32(float32x2_t a, float32x2_t b, float32_t c)
 {
     float32x2_t res;
@@ -14175,7 +14177,7 @@ _NEON2SSE_INLINE float32x2_t vmls_n_f32(float32x2_t a, float32x2_t b, float32_t
     return res;
 }
 
-int16x8_t vmlsq_n_s16(int16x8_t a, int16x8_t b, int16_t c); // VMLS.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE int16x8_t vmlsq_n_s16(int16x8_t a, int16x8_t b, int16_t c); // VMLS.I16 q0, q0, d0[0]
 _NEON2SSE_INLINE int16x8_t vmlsq_n_s16(int16x8_t a, int16x8_t b, int16_t c) // VMLS.I16 q0, q0, d0[0]
 {
     int16x8_t vc;
@@ -14183,7 +14185,7 @@ _NEON2SSE_INLINE int16x8_t vmlsq_n_s16(int16x8_t a, int16x8_t b, int16_t c) // V
     return vmlsq_s16(a, b,vc);
 }
 
-int32x4_t vmlsq_n_s32(int32x4_t a, int32x4_t b, int32_t c); // VMLS.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlsq_n_s32(int32x4_t a, int32x4_t b, int32_t c); // VMLS.I32 q0, q0, d0[0]
 _NEON2SSE_INLINE int32x4_t vmlsq_n_s32(int32x4_t a, int32x4_t b, int32_t c) // VMLS.I32 q0, q0, d0[0]
 {
     int32x4_t vc;
@@ -14191,15 +14193,15 @@ _NEON2SSE_INLINE int32x4_t vmlsq_n_s32(int32x4_t a, int32x4_t b, int32_t c) // V
     return vmlsq_s32(a,b,vc);
 }
 
-uint16x8_t vmlsq_n_u16(uint16x8_t a, uint16x8_t b, uint16_t c); // VMLS.I16 q0, q0, d0[0]
+_NEON2SSESTORAGE uint16x8_t vmlsq_n_u16(uint16x8_t a, uint16x8_t b, uint16_t c); // VMLS.I16 q0, q0, d0[0]
 _NEON2SSE_INLINE uint16x8_t vmlsq_n_u16(uint16x8_t a, uint16x8_t b, uint16_t c) // VMLS.I16 q0, q0, d0[0]
 {
-    uint16x8_t vc; 
-    vc = vdupq_n_u16(c); 
-    return vmlsq_u16(a,b,vc); 
+    uint16x8_t vc;
+    vc = vdupq_n_u16(c);
+    return vmlsq_u16(a,b,vc);
 }
 
-uint32x4_t vmlsq_n_u32(uint32x4_t a, uint32x4_t b, uint32_t c); // VMLS.I32 q0, q0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlsq_n_u32(uint32x4_t a, uint32x4_t b, uint32_t c); // VMLS.I32 q0, q0, d0[0]
 _NEON2SSE_INLINE uint32x4_t vmlsq_n_u32(uint32x4_t a, uint32x4_t b, uint32_t c) // VMLS.I32 q0, q0, d0[0]
 {
     uint32x4_t vc;
@@ -14207,7 +14209,7 @@ _NEON2SSE_INLINE uint32x4_t vmlsq_n_u32(uint32x4_t a, uint32x4_t b, uint32_t c)
     return vmlsq_u32(a,b,vc);
 }
 
-float32x4_t vmlsq_n_f32(float32x4_t a, float32x4_t b, float32_t c); // VMLS.F32 q0, q0, d0[0]
+_NEON2SSESTORAGE float32x4_t vmlsq_n_f32(float32x4_t a, float32x4_t b, float32_t c); // VMLS.F32 q0, q0, d0[0]
 _NEON2SSE_INLINE float32x4_t vmlsq_n_f32(float32x4_t a, float32x4_t b, float32_t c)
 {
     float32x4_t vc;
@@ -14216,7 +14218,7 @@ _NEON2SSE_INLINE float32x4_t vmlsq_n_f32(float32x4_t a, float32x4_t b, float32_t
 }
 
 //**** Vector widening multiply subtract with scalar ******
-int32x4_t vmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VMLSL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VMLSL.S16 q0, d0, d0[0]
 _NEON2SSE_INLINE int32x4_t vmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c) // VMLSL.S16 q0, d0, d0[0]
 {
     int16x4_t vc;
@@ -14224,7 +14226,7 @@ _NEON2SSE_INLINE int32x4_t vmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c) // V
     return vmlsl_s16(a, b, vc);
 }
 
-int64x2_t vmlsl_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VMLSL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vmlsl_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VMLSL.S32 q0, d0, d0[0]
 _NEON2SSE_INLINE int64x2_t vmlsl_n_s32(int64x2_t a, int32x2_t b, int32_t c) // VMLSL.S32 q0, d0, d0[0]
 {
     int32x2_t vc;
@@ -14232,7 +14234,7 @@ _NEON2SSE_INLINE int64x2_t vmlsl_n_s32(int64x2_t a, int32x2_t b, int32_t c) // V
     return vmlsl_s32(a, b, vc);
 }
 
-uint32x4_t vmlsl_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c); // VMLSL.s16 q0, d0, d0[0]
+_NEON2SSESTORAGE uint32x4_t vmlsl_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c); // VMLSL.s16 q0, d0, d0[0]
 _NEON2SSE_INLINE uint32x4_t vmlsl_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c) // VMLSL.s16 q0, d0, d0[0]
 {
     uint16x4_t vc;
@@ -14240,7 +14242,7 @@ _NEON2SSE_INLINE uint32x4_t vmlsl_n_u16(uint32x4_t a, uint16x4_t b, uint16_t c)
     return vmlsl_u16(a, b, vc);
 }
 
-uint64x2_t vmlsl_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c); // VMLSL.U32 q0, d0, d0[0]
+_NEON2SSESTORAGE uint64x2_t vmlsl_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c); // VMLSL.U32 q0, d0, d0[0]
 _NEON2SSE_INLINE uint64x2_t vmlsl_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c) // VMLSL.U32 q0, d0, d0[0]
 {
     uint32x2_t vc;
@@ -14250,7 +14252,7 @@ _NEON2SSE_INLINE uint64x2_t vmlsl_n_u32(uint64x2_t a, uint32x2_t b, uint32_t c)
 
 //***** Vector widening saturating doubling multiply subtract with scalar *********
 //**********************************************************************************
-int32x4_t vqdmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VQDMLSL.S16 q0, d0, d0[0]
+_NEON2SSESTORAGE int32x4_t vqdmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c); // VQDMLSL.S16 q0, d0, d0[0]
 _NEON2SSE_INLINE int32x4_t vqdmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c)
 {
     int16x4_t vc;
@@ -14258,7 +14260,7 @@ _NEON2SSE_INLINE int32x4_t vqdmlsl_n_s16(int32x4_t a, int16x4_t b, int16_t c)
     return vqdmlsl_s16(a, b, vc);
 }
 
-int64x2_t vqdmlsl_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VQDMLSL.S32 q0, d0, d0[0]
+_NEON2SSESTORAGE int64x2_t vqdmlsl_n_s32(int64x2_t a, int32x2_t b, int32_t c); // VQDMLSL.S32 q0, d0, d0[0]
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmlsl_n_s32(int64x2_t a, int32x2_t b, int32_t c), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int32x2_t vc;
@@ -14271,7 +14273,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int64x2_t vqdmlsl_n_s32(int64x2_t
 //VEXT (Vector Extract) extracts  elements from the bottom end of the second operand
 //vector and the top end of the first, concatenates them, and places the result in the destination vector
 //c elements from the bottom end of the second operand and (8-c) from the top end of the first
-int8x8_t vext_s8(int8x8_t a, int8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
+_NEON2SSESTORAGE int8x8_t vext_s8(int8x8_t a, int8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x8_t vext_s8(int8x8_t a, int8x8_t b, __constrange(0,7) int c),_NEON2SSE_REASON_SLOW_SERIAL)
 {
     int8x8_t res;
@@ -14285,14 +14287,14 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int8x8_t vext_s8(int8x8_t a, int8
     return res;
 }
 
-uint8x8_t vext_u8(uint8x8_t a,  uint8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
+_NEON2SSESTORAGE uint8x8_t vext_u8(uint8x8_t a,  uint8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
 #define vext_u8 vext_s8
 //same result tested
 
-poly8x8_t vext_p8(poly8x8_t a, poly8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
+_NEON2SSESTORAGE poly8x8_t vext_p8(poly8x8_t a, poly8x8_t b, __constrange(0,7) int c); // VEXT.8 d0,d0,d0,#0
 #define vext_p8 vext_u8
 
-int16x4_t vext_s16(int16x4_t a, int16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
+_NEON2SSESTORAGE int16x4_t vext_s16(int16x4_t a, int16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
 _NEON2SSE_INLINE int16x4_t  _NEON2SSE_PERFORMANCE_WARNING (vext_s16(int16x4_t a, int16x4_t b, __constrange(0,3) int c), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int16x4_t res;
@@ -14306,13 +14308,13 @@ _NEON2SSE_INLINE int16x4_t  _NEON2SSE_PERFORMANCE_WARNING (vext_s16(int16x4_t a,
     return res;
 }
 
-uint16x4_t vext_u16(uint16x4_t a,  uint16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
+_NEON2SSESTORAGE uint16x4_t vext_u16(uint16x4_t a,  uint16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
 #define vext_u16 vext_s16
 
-poly16x4_t vext_p16(poly16x4_t a, poly16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
+_NEON2SSESTORAGE poly16x4_t vext_p16(poly16x4_t a, poly16x4_t b, __constrange(0,3) int c); // VEXT.16 d0,d0,d0,#0
 #define vext_p16 vext_s16
 
-int32x2_t vext_s32(int32x2_t a, int32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
+_NEON2SSESTORAGE int32x2_t vext_s32(int32x2_t a, int32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vext_s32(int32x2_t a, int32x2_t b, __constrange(0,1) int c), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     int32x2_t res;
@@ -14326,7 +14328,7 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(int32x2_t vext_s32(int32x2_t a, i
     return res;
 }
 
-float32x2_t vext_f32(float32x2_t a, float32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
+_NEON2SSESTORAGE float32x2_t vext_f32(float32x2_t a, float32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(float32x2_t vext_f32(float32x2_t a, float32x2_t b, __constrange(0,1) int c), _NEON2SSE_REASON_SLOW_SERIAL)
 {
     float32x2_t res;
@@ -14340,53 +14342,53 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(float32x2_t vext_f32(float32x2_t
     return res;
 }
 
-uint32x2_t vext_u32(uint32x2_t a,  uint32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
+_NEON2SSESTORAGE uint32x2_t vext_u32(uint32x2_t a,  uint32x2_t b, __constrange(0,1) int c); // VEXT.32 d0,d0,d0,#0
 #define vext_u32 vext_s32
 
 
-int64x1_t vext_s64(int64x1_t a, int64x1_t b, __constrange(0,0) int c); // VEXT.64 d0,d0,d0,#0
+_NEON2SSESTORAGE int64x1_t vext_s64(int64x1_t a, int64x1_t b, __constrange(0,0) int c); // VEXT.64 d0,d0,d0,#0
 #define vext_s64(a,b,c) a
 
-uint64x1_t vext_u64(uint64x1_t a, uint64x1_t b, __constrange(0,0) int c); // VEXT.64 d0,d0,d0,#0
+_NEON2SSESTORAGE uint64x1_t vext_u64(uint64x1_t a, uint64x1_t b, __constrange(0,0) int c); // VEXT.64 d0,d0,d0,#0
 #define vext_u64(a,b,c) a
 
-int8x16_t vextq_s8(int8x16_t a, int8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
+_NEON2SSESTORAGE int8x16_t vextq_s8(int8x16_t a, int8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
 #define vextq_s8(a,b,c) _MM_ALIGNR_EPI8 (b,a,c)
 
-uint8x16_t vextq_u8(uint8x16_t a, uint8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
+_NEON2SSESTORAGE uint8x16_t vextq_u8(uint8x16_t a, uint8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
 #define vextq_u8(a,b,c) _MM_ALIGNR_EPI8 (b,a,c)
 
-poly8x16_t vextq_p8(poly8x16_t a, poly8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
+_NEON2SSESTORAGE poly8x16_t vextq_p8(poly8x16_t a, poly8x16_t b, __constrange(0,15) int c); // VEXT.8 q0,q0,q0,#0
 #define vextq_p8 vextq_s8
 
-int16x8_t vextq_s16(int16x8_t a, int16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
+_NEON2SSESTORAGE int16x8_t vextq_s16(int16x8_t a, int16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
 #define vextq_s16(a,b,c) _MM_ALIGNR_EPI8 (b,a,c * 2)
 
-uint16x8_t vextq_u16(uint16x8_t a, uint16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
+_NEON2SSESTORAGE uint16x8_t vextq_u16(uint16x8_t a, uint16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
 #define vextq_u16(a,b,c) _MM_ALIGNR_EPI8 (b,a,c * 2)
 
-poly16x8_t vextq_p16(poly16x8_t a, poly16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
+_NEON2SSESTORAGE poly16x8_t vextq_p16(poly16x8_t a, poly16x8_t b, __constrange(0,7) int c); // VEXT.16 q0,q0,q0,#0
 #define vextq_p16 vextq_s16
 
-int32x4_t vextq_s32(int32x4_t a, int32x4_t b, __constrange(0,3) int c); // VEXT.32 q0,q0,q0,#0
+_NEON2SSESTORAGE int32x4_t vextq_s32(int32x4_t a, int32x4_t b, __constrange(0,3) int c); // VEXT.32 q0,q0,q0,#0
 #define vextq_s32(a,b,c) _MM_ALIGNR_EPI8 (b,a,c * 4)
 
-uint32x4_t vextq_u32(uint32x4_t a, uint32x4_t b, __constrange(0,3) int c); // VEXT.32 q0,q0,q0,#0
+_NEON2SSESTORAGE uint32x4_t vextq_u32(uint32x4_t a, uint32x4_t b, __constrange(0,3) int c); // VEXT.32 q0,q0,q0,#0
 #define vextq_u32(a,b,c) _MM_ALIGNR_EPI8 (b,a,c * 4)
 
-float32x4_t vextq_f32(float32x4_t a, float32x4_t b, __constrange(0,3) float c); // VEXT.32 q0,q0,q0,#0
+_NEON2SSESTORAGE float32x4_t vextq_f32(float32x4_t a, float32x4_t b, __constrange(0,3) float c); // VEXT.32 q0,q0,q0,#0
 #define vextq_f32(a,b,c) _M128(vextq_s32(_M128i(a),_M128i(b),c) )
 
-int64x2_t vextq_s64(int64x2_t a, int64x2_t b, __constrange(0,1) int c); // VEXT.64 q0,q0,q0,#0
+_NEON2SSESTORAGE int64x2_t vextq_s64(int64x2_t a, int64x2_t b, __constrange(0,1) int c); // VEXT.64 q0,q0,q0,#0
 #define vextq_s64(a,b,c) _MM_ALIGNR_EPI8(b,a,c * 8)
 
-uint64x2_t vextq_u64(uint64x2_t a, uint64x2_t b, __constrange(0,1) int c); // VEXT.64 q0,q0,q0,#0
+_NEON2SSESTORAGE uint64x2_t vextq_u64(uint64x2_t a, uint64x2_t b, __constrange(0,1) int c); // VEXT.64 q0,q0,q0,#0
 #define vextq_u64(a,b,c) _MM_ALIGNR_EPI8(b,a,c * 8)
 
 //************ Reverse vector elements (swap endianness)*****************
 //*************************************************************************
 //VREVn.m reverses the order of the m-bit lanes within a set that is n bits wide.
-int8x8_t vrev64_s8(int8x8_t vec); // VREV64.8 d0,d0
+_NEON2SSESTORAGE int8x8_t vrev64_s8(int8x8_t vec); // VREV64.8 d0,d0
 _NEON2SSE_INLINE int8x8_t vrev64_s8(int8x8_t vec)
 {
     int8x8_t res64;
@@ -14395,7 +14397,7 @@ _NEON2SSE_INLINE int8x8_t vrev64_s8(int8x8_t vec)
     return64(res);
 }
 
-int16x4_t vrev64_s16(int16x4_t vec); // VREV64.16 d0,d0
+_NEON2SSESTORAGE int16x4_t vrev64_s16(int16x4_t vec); // VREV64.16 d0,d0
 _NEON2SSE_INLINE int16x4_t vrev64_s16(int16x4_t vec)
 {
     int16x4_t res64;
@@ -14404,7 +14406,7 @@ _NEON2SSE_INLINE int16x4_t vrev64_s16(int16x4_t vec)
     return64(res);
 }
 
-int32x2_t vrev64_s32(int32x2_t vec); // VREV64.32 d0,d0
+_NEON2SSESTORAGE int32x2_t vrev64_s32(int32x2_t vec); // VREV64.32 d0,d0
 _NEON2SSE_INLINE int32x2_t vrev64_s32(int32x2_t vec)
 {
     int32x2_t res;
@@ -14413,22 +14415,22 @@ _NEON2SSE_INLINE int32x2_t vrev64_s32(int32x2_t vec)
     return res;
 }
 
-uint8x8_t vrev64_u8(uint8x8_t vec); // VREV64.8 d0,d0
+_NEON2SSESTORAGE uint8x8_t vrev64_u8(uint8x8_t vec); // VREV64.8 d0,d0
 #define vrev64_u8 vrev64_s8
 
-uint16x4_t vrev64_u16(uint16x4_t vec); // VREV64.16 d0,d0
+_NEON2SSESTORAGE uint16x4_t vrev64_u16(uint16x4_t vec); // VREV64.16 d0,d0
 #define vrev64_u16 vrev64_s16
 
-uint32x2_t vrev64_u32(uint32x2_t vec); // VREV64.32 d0,d0
+_NEON2SSESTORAGE uint32x2_t vrev64_u32(uint32x2_t vec); // VREV64.32 d0,d0
 #define vrev64_u32 vrev64_s32
 
-poly8x8_t vrev64_p8(poly8x8_t vec); // VREV64.8 d0,d0
+_NEON2SSESTORAGE poly8x8_t vrev64_p8(poly8x8_t vec); // VREV64.8 d0,d0
 #define vrev64_p8 vrev64_u8
 
-poly16x4_t vrev64_p16(poly16x4_t vec); // VREV64.16 d0,d0
+_NEON2SSESTORAGE poly16x4_t vrev64_p16(poly16x4_t vec); // VREV64.16 d0,d0
 #define vrev64_p16 vrev64_u16
 
-float32x2_t vrev64_f32(float32x2_t vec); // VREV64.32 d0,d0
+_NEON2SSESTORAGE float32x2_t vrev64_f32(float32x2_t vec); // VREV64.32 d0,d0
 _NEON2SSE_INLINE float32x2_t vrev64_f32(float32x2_t vec)
 {
     float32x2_t res;
@@ -14437,48 +14439,48 @@ _NEON2SSE_INLINE float32x2_t vrev64_f32(float32x2_t vec)
     return res;
 }
 
-int8x16_t vrev64q_s8(int8x16_t vec); // VREV64.8 q0,q0
+_NEON2SSESTORAGE int8x16_t vrev64q_s8(int8x16_t vec); // VREV64.8 q0,q0
 _NEON2SSE_INLINE int8x16_t vrev64q_s8(int8x16_t vec) // VREV64.8 q0,q0
 {
-    _NEON2SSE_ALIGN_16 int8_t mask_rev_e8[16] = {7,6,5,4,3,2,1,0, 15,14,13,12,11,10,9, 8};
+    _NEON2SSE_ALIGN_16 static const int8_t mask_rev_e8[16] = {7,6,5,4,3,2,1,0, 15,14,13,12,11,10,9, 8};
     return _mm_shuffle_epi8 (vec, *(__m128i*)  mask_rev_e8);
 }
 
-int16x8_t vrev64q_s16(int16x8_t vec); // VREV64.16 q0,q0
+_NEON2SSESTORAGE int16x8_t vrev64q_s16(int16x8_t vec); // VREV64.16 q0,q0
 _NEON2SSE_INLINE int16x8_t vrev64q_s16(int16x8_t vec) // VREV64.16 q0,q0
 {
     //no _mm_shuffle_epi16, _mm_shuffle_epi8 to be used with the corresponding mask
-    _NEON2SSE_ALIGN_16 int8_t mask_rev_e16[16] = {6,7, 4,5,2,3,0,1,14,15,12,13,10,11,8,9};
+    _NEON2SSE_ALIGN_16 static const int8_t mask_rev_e16[16] = {6,7, 4,5,2,3,0,1,14,15,12,13,10,11,8,9};
     return _mm_shuffle_epi8 (vec, *(__m128i*)mask_rev_e16);
 }
 
-int32x4_t vrev64q_s32(int32x4_t vec); // VREV64.32 q0,q0
+_NEON2SSESTORAGE int32x4_t vrev64q_s32(int32x4_t vec); // VREV64.32 q0,q0
 _NEON2SSE_INLINE int32x4_t vrev64q_s32(int32x4_t vec) // VREV64.32 q0,q0
 {
     return _mm_shuffle_epi32 (vec, 1 | (0 << 2) | (3 << 4) | (2 << 6) );
 }
 
-uint8x16_t vrev64q_u8(uint8x16_t vec); // VREV64.8 q0,q0
+_NEON2SSESTORAGE uint8x16_t vrev64q_u8(uint8x16_t vec); // VREV64.8 q0,q0
 #define vrev64q_u8 vrev64q_s8
 
-uint16x8_t vrev64q_u16(uint16x8_t vec); // VREV64.16 q0,q0
+_NEON2SSESTORAGE uint16x8_t vrev64q_u16(uint16x8_t vec); // VREV64.16 q0,q0
 #define vrev64q_u16 vrev64q_s16
 
-uint32x4_t vrev64q_u32(uint32x4_t vec); // VREV64.32 q0,q0
+_NEON2SSESTORAGE uint32x4_t vrev64q_u32(uint32x4_t vec); // VREV64.32 q0,q0
 #define vrev64q_u32 vrev64q_s32
 
-poly8x16_t vrev64q_p8(poly8x16_t vec); // VREV64.8 q0,q0
+_NEON2SSESTORAGE poly8x16_t vrev64q_p8(poly8x16_t vec); // VREV64.8 q0,q0
 #define vrev64q_p8 vrev64q_u8
 
-poly16x8_t vrev64q_p16(poly16x8_t vec); // VREV64.16 q0,q0
+_NEON2SSESTORAGE poly16x8_t vrev64q_p16(poly16x8_t vec); // VREV64.16 q0,q0
 #define vrev64q_p16 vrev64q_u16
 
-float32x4_t vrev64q_f32(float32x4_t vec); // VREV64.32 q0,q0
+_NEON2SSESTORAGE float32x4_t vrev64q_f32(float32x4_t vec); // VREV64.32 q0,q0
 #define vrev64q_f32(vec) _mm_shuffle_ps (vec,  vec, _MM_SHUFFLE(2,3, 0,1))
 
 //********************  32 bit shuffles **********************
 //************************************************************
-int8x8_t vrev32_s8(int8x8_t vec); // VREV32.8 d0,d0
+_NEON2SSESTORAGE int8x8_t vrev32_s8(int8x8_t vec); // VREV32.8 d0,d0
 _NEON2SSE_INLINE int8x8_t vrev32_s8(int8x8_t vec)
 {
     int8x8_t res64;
@@ -14487,7 +14489,7 @@ _NEON2SSE_INLINE int8x8_t vrev32_s8(int8x8_t vec)
     return64(res);
 }
 
-int16x4_t vrev32_s16(int16x4_t vec); // VREV32.16 d0,d0
+_NEON2SSESTORAGE int16x4_t vrev32_s16(int16x4_t vec); // VREV32.16 d0,d0
 _NEON2SSE_INLINE int16x4_t vrev32_s16(int16x4_t vec)
 {
     int16x4_t res64;
@@ -14496,47 +14498,47 @@ _NEON2SSE_INLINE int16x4_t vrev32_s16(int16x4_t vec)
     return64(res);
 }
 
-uint8x8_t vrev32_u8(uint8x8_t vec); // VREV32.8 d0,d0
+_NEON2SSESTORAGE uint8x8_t vrev32_u8(uint8x8_t vec); // VREV32.8 d0,d0
 #define vrev32_u8 vrev32_s8
 
-uint16x4_t vrev32_u16(uint16x4_t vec); // VREV32.16 d0,d0
+_NEON2SSESTORAGE uint16x4_t vrev32_u16(uint16x4_t vec); // VREV32.16 d0,d0
 #define vrev32_u16 vrev32_s16
 
-poly8x8_t vrev32_p8(poly8x8_t vec); // VREV32.8 d0,d0
+_NEON2SSESTORAGE poly8x8_t vrev32_p8(poly8x8_t vec); // VREV32.8 d0,d0
 #define vrev32_p8 vrev32_u8
 
-poly16x4_t vrev32_p16(poly16x4_t vec); // VREV32.16 d0,d0
+_NEON2SSESTORAGE poly16x4_t vrev32_p16(poly16x4_t vec); // VREV32.16 d0,d0
 #define vrev32_p16 vrev32_u16
 
-int8x16_t vrev32q_s8(int8x16_t vec); // VREV32.8 q0,q0
+_NEON2SSESTORAGE int8x16_t vrev32q_s8(int8x16_t vec); // VREV32.8 q0,q0
 _NEON2SSE_INLINE int8x16_t vrev32q_s8(int8x16_t vec) // VREV32.8 q0,q0
 {
-    _NEON2SSE_ALIGN_16 int8_t mask_rev_e8[16] = {3,2,1,0, 7,6,5,4, 11,10,9,8, 15,14,13,12};
+    _NEON2SSE_ALIGN_16 static const int8_t mask_rev_e8[16] = {3,2,1,0, 7,6,5,4, 11,10,9,8, 15,14,13,12};
     return _mm_shuffle_epi8 (vec, *(__m128i*)  mask_rev_e8);
 }
 
-int16x8_t vrev32q_s16(int16x8_t vec); // VREV32.16 q0,q0
+_NEON2SSESTORAGE int16x8_t vrev32q_s16(int16x8_t vec); // VREV32.16 q0,q0
 _NEON2SSE_INLINE int16x8_t vrev32q_s16(int16x8_t vec) // VREV32.16 q0,q0
 {
-    _NEON2SSE_ALIGN_16 int8_t mask_rev_e8[16] = {2,3,0,1, 6,7, 4,5, 10,11, 8,9, 14,15,12,13};
+    _NEON2SSE_ALIGN_16 static const int8_t mask_rev_e8[16] = {2,3,0,1, 6,7, 4,5, 10,11, 8,9, 14,15,12,13};
     return _mm_shuffle_epi8 (vec, *(__m128i*)  mask_rev_e8);
 }
 
-uint8x16_t vrev32q_u8(uint8x16_t vec); // VREV32.8 q0,q0
+_NEON2SSESTORAGE uint8x16_t vrev32q_u8(uint8x16_t vec); // VREV32.8 q0,q0
 #define vrev32q_u8 vrev32q_s8
 
-uint16x8_t vrev32q_u16(uint16x8_t vec); // VREV32.16 q0,q0
+_NEON2SSESTORAGE uint16x8_t vrev32q_u16(uint16x8_t vec); // VREV32.16 q0,q0
 #define vrev32q_u16 vrev32q_s16
 
-poly8x16_t vrev32q_p8(poly8x16_t vec); // VREV32.8 q0,q0
+_NEON2SSESTORAGE poly8x16_t vrev32q_p8(poly8x16_t vec); // VREV32.8 q0,q0
 #define vrev32q_p8 vrev32q_u8
 
-poly16x8_t vrev32q_p16(poly16x8_t vec); // VREV32.16 q0,q0
+_NEON2SSESTORAGE poly16x8_t vrev32q_p16(poly16x8_t vec); // VREV32.16 q0,q0
 #define vrev32q_p16 vrev32q_u16
 
 //*************  16 bit shuffles **********************
 //******************************************************
-int8x8_t vrev16_s8(int8x8_t vec); // VREV16.8 d0,d0
+_NEON2SSESTORAGE int8x8_t vrev16_s8(int8x8_t vec); // VREV16.8 d0,d0
 _NEON2SSE_INLINE int8x8_t vrev16_s8(int8x8_t vec)
 {
     int8x8_t res64;
@@ -14545,23 +14547,23 @@ _NEON2SSE_INLINE int8x8_t vrev16_s8(int8x8_t vec)
     return64(res);
 }
 
-uint8x8_t vrev16_u8(uint8x8_t vec); // VREV16.8 d0,d0
+_NEON2SSESTORAGE uint8x8_t vrev16_u8(uint8x8_t vec); // VREV16.8 d0,d0
 #define vrev16_u8 vrev16_s8
 
-poly8x8_t vrev16_p8(poly8x8_t vec); // VREV16.8 d0,d0
+_NEON2SSESTORAGE poly8x8_t vrev16_p8(poly8x8_t vec); // VREV16.8 d0,d0
 #define vrev16_p8 vrev16_u8
 
-int8x16_t vrev16q_s8(int8x16_t vec); // VREV16.8 q0,q0
+_NEON2SSESTORAGE int8x16_t vrev16q_s8(int8x16_t vec); // VREV16.8 q0,q0
 _NEON2SSE_INLINE int8x16_t vrev16q_s8(int8x16_t vec) // VREV16.8 q0,q0
 {
-    _NEON2SSE_ALIGN_16 int8_t mask_rev8[16] = {1,0, 3,2, 5,4, 7,6, 9,8, 11, 10, 13, 12, 15, 14};
+    _NEON2SSE_ALIGN_16 static const int8_t mask_rev8[16] = {1,0, 3,2, 5,4, 7,6, 9,8, 11, 10, 13, 12, 15, 14};
     return _mm_shuffle_epi8 (vec, *(__m128i*)  mask_rev8);
 }
 
-uint8x16_t vrev16q_u8(uint8x16_t vec); // VREV16.8 q0,q0
+_NEON2SSESTORAGE uint8x16_t vrev16q_u8(uint8x16_t vec); // VREV16.8 q0,q0
 #define vrev16q_u8 vrev16q_s8
 
-poly8x16_t vrev16q_p8(poly8x16_t vec); // VREV16.8 q0,q0
+_NEON2SSESTORAGE poly8x16_t vrev16q_p8(poly8x16_t vec); // VREV16.8 q0,q0
 #define vrev16q_p8 vrev16q_u8
 
 //*********************************************************************
@@ -14570,7 +14572,7 @@ poly8x16_t vrev16q_p8(poly8x16_t vec); // VREV16.8 q0,q0
 
 //*********** Absolute: Vd[i] = |Va[i]| **********************************
 //************************************************************************
-int8x8_t   vabs_s8(int8x8_t a); // VABS.S8 d0,d0
+_NEON2SSESTORAGE int8x8_t   vabs_s8(int8x8_t a); // VABS.S8 d0,d0
 _NEON2SSE_INLINE int8x8_t   vabs_s8(int8x8_t a)
 {
     int8x8_t res64;
@@ -14580,7 +14582,7 @@ _NEON2SSE_INLINE int8x8_t   vabs_s8(int8x8_t a)
 }
 
 
-int16x4_t   vabs_s16(int16x4_t a); // VABS.S16 d0,d0
+_NEON2SSESTORAGE int16x4_t   vabs_s16(int16x4_t a); // VABS.S16 d0,d0
 _NEON2SSE_INLINE int16x4_t   vabs_s16(int16x4_t a)
 {
     int16x4_t res64;
@@ -14589,7 +14591,7 @@ _NEON2SSE_INLINE int16x4_t   vabs_s16(int16x4_t a)
     return64(res);
 }
 
-int32x2_t   vabs_s32(int32x2_t a); // VABS.S32 d0,d0
+_NEON2SSESTORAGE int32x2_t   vabs_s32(int32x2_t a); // VABS.S32 d0,d0
 _NEON2SSE_INLINE int32x2_t   vabs_s32(int32x2_t a)
 {
     int32x2_t res64;
@@ -14598,45 +14600,45 @@ _NEON2SSE_INLINE int32x2_t   vabs_s32(int32x2_t a)
     return64(res);
 }
 
-float32x2_t vabs_f32(float32x2_t a); // VABS.F32 d0,d0
+_NEON2SSESTORAGE float32x2_t vabs_f32(float32x2_t a); // VABS.F32 d0,d0
 _NEON2SSE_INLINE float32x2_t vabs_f32(float32x2_t a) // VABS.F32 d0,d0
 {
     float32x4_t res;
     __m64_128 res64;
-    _NEON2SSE_ALIGN_16 int32_t c7fffffff[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff};
+    _NEON2SSE_ALIGN_16 static const int32_t c7fffffff[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff};
     res = _mm_and_ps (_pM128(a), *(__m128*)c7fffffff); //use 64 low bits only
     _M64f(res64, res);
     return res64;
 }
 
-int8x16_t   vabsq_s8(int8x16_t a); // VABS.S8 q0,q0
+_NEON2SSESTORAGE int8x16_t   vabsq_s8(int8x16_t a); // VABS.S8 q0,q0
 #define vabsq_s8 _mm_abs_epi8
 
-int16x8_t   vabsq_s16(int16x8_t a); // VABS.S16 q0,q0
+_NEON2SSESTORAGE int16x8_t   vabsq_s16(int16x8_t a); // VABS.S16 q0,q0
 #define vabsq_s16 _mm_abs_epi16
 
-int32x4_t   vabsq_s32(int32x4_t a); // VABS.S32 q0,q0
+_NEON2SSESTORAGE int32x4_t   vabsq_s32(int32x4_t a); // VABS.S32 q0,q0
 #define vabsq_s32 _mm_abs_epi32
 
-float32x4_t vabsq_f32(float32x4_t a); // VABS.F32 q0,q0
+_NEON2SSESTORAGE float32x4_t vabsq_f32(float32x4_t a); // VABS.F32 q0,q0
 _NEON2SSE_INLINE float32x4_t vabsq_f32(float32x4_t a) // VABS.F32 q0,q0
 {
-    _NEON2SSE_ALIGN_16 int32_t c7fffffff[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff};
+    _NEON2SSE_ALIGN_16 static const int32_t c7fffffff[4] = {0x7fffffff, 0x7fffffff, 0x7fffffff, 0x7fffffff};
     return _mm_and_ps (a, *(__m128*)c7fffffff);
 }
 
 #ifdef _NEON2SSE_64BIT
-int64x2_t vabsq_s64(int64x2_t a); // VABS.S64 q0,q0
+_NEON2SSESTORAGE int64x2_t vabsq_s64(int64x2_t a); // VABS.S64 q0,q0
 _NEON2SSE_INLINE int64x2_t vabsq_s64(int64x2_t a) // VABS.S64 q0,q0
 {
     __m128i sign = _mm_srai_epi32 (_mm_shuffle_epi32 (a, 0xf5), 31);
     return _mm_sub_epi64 (_mm_xor_si128 (a, sign), sign);
 }
 
-float64x2_t vabsq_f64(float64x2_t a); // VABS.F64 q0,q0
+_NEON2SSESTORAGE float64x2_t vabsq_f64(float64x2_t a); // VABS.F64 q0,q0
 _NEON2SSE_INLINE float64x2_t vabsq_f64(float64x2_t a) // VABS.F64 q0,q0
 {
-    _NEON2SSE_ALIGN_16 int64_t mask[2] = {0x7fffffffffffffffLL, 0x7fffffffffffffffLL};
+    _NEON2SSE_ALIGN_16 static const int64_t mask[2] = {0x7fffffffffffffffLL, 0x7fffffffffffffffLL};
     return _mm_and_pd (a, *(__m128d*)mask);
 }
 #endif
@@ -14644,7 +14646,7 @@ _NEON2SSE_INLINE float64x2_t vabsq_f64(float64x2_t a) // VABS.F64 q0,q0
 //****** Saturating absolute: Vd[i] = sat(|Va[i]|) *********************
 //**********************************************************************
 //For signed-integer data types, the absolute value of the most negative value is not representable by the data type, saturation takes place
-int8x8_t vqabs_s8(int8x8_t a); // VQABS.S8 d0,d0
+_NEON2SSESTORAGE int8x8_t vqabs_s8(int8x8_t a); // VQABS.S8 d0,d0
 _NEON2SSE_INLINE int8x8_t vqabs_s8(int8x8_t a)
 {
     int8x8_t res64;
@@ -14653,7 +14655,7 @@ _NEON2SSE_INLINE int8x8_t vqabs_s8(int8x8_t a)
     return64(res);
 }
 
-int16x4_t vqabs_s16(int16x4_t a); // VQABS.S16 d0,d0
+_NEON2SSESTORAGE int16x4_t vqabs_s16(int16x4_t a); // VQABS.S16 d0,d0
 _NEON2SSE_INLINE int16x4_t vqabs_s16(int16x4_t a)
 {
     int16x4_t res64;
@@ -14662,7 +14664,7 @@ _NEON2SSE_INLINE int16x4_t vqabs_s16(int16x4_t a)
     return64(res);
 }
 
-int32x2_t vqabs_s32(int32x2_t a); // VQABS.S32 d0,d0
+_NEON2SSESTORAGE int32x2_t vqabs_s32(int32x2_t a); // VQABS.S32 d0,d0
 _NEON2SSE_INLINE int32x2_t vqabs_s32(int32x2_t a)
 {
     int32x2_t res64;
@@ -14671,7 +14673,7 @@ _NEON2SSE_INLINE int32x2_t vqabs_s32(int32x2_t a)
     return64(res);
 }
 
-int8x16_t vqabsq_s8(int8x16_t a); // VQABS.S8 q0,q0
+_NEON2SSESTORAGE int8x16_t vqabsq_s8(int8x16_t a); // VQABS.S8 q0,q0
 _NEON2SSE_INLINE int8x16_t vqabsq_s8(int8x16_t a) // VQABS.S8 q0,q0
 {
     __m128i c_128, abs, abs_cmp;
@@ -14681,7 +14683,7 @@ _NEON2SSE_INLINE int8x16_t vqabsq_s8(int8x16_t a) // VQABS.S8 q0,q0
     return _mm_xor_si128 (abs,  abs_cmp);
 }
 
-int16x8_t vqabsq_s16(int16x8_t a); // VQABS.S16 q0,q0
+_NEON2SSESTORAGE int16x8_t vqabsq_s16(int16x8_t a); // VQABS.S16 q0,q0
 _NEON2SSE_INLINE int16x8_t vqabsq_s16(int16x8_t a) // VQABS.S16 q0,q0
 {
     __m128i c_32768, abs, abs_cmp;
@@ -14691,7 +14693,7 @@ _NEON2SSE_INLINE int16x8_t vqabsq_s16(int16x8_t a) // VQABS.S16 q0,q0
     return _mm_xor_si128 (abs,  abs_cmp);
 }
 
-int32x4_t vqabsq_s32(int32x4_t a); // VQABS.S32 q0,q0
+_NEON2SSESTORAGE int32x4_t vqabsq_s32(int32x4_t a); // VQABS.S32 q0,q0
 _NEON2SSE_INLINE int32x4_t vqabsq_s32(int32x4_t a) // VQABS.S32 q0,q0
 {
     __m128i c80000000, abs, abs_cmp;
@@ -14705,7 +14707,7 @@ _NEON2SSE_INLINE int32x4_t vqabsq_s32(int32x4_t a) // VQABS.S32 q0,q0
 //*****************************************************************************
 //several Negate implementations possible for SIMD.
 //e.//function _mm_sign function(a, negative numbers vector), but the following one gives good performance:
-int8x8_t vneg_s8(int8x8_t a); // VNE//d0,d0
+_NEON2SSESTORAGE int8x8_t vneg_s8(int8x8_t a); // VNE//d0,d0
 _NEON2SSE_INLINE int8x8_t vneg_s8(int8x8_t a)
 {
     int8x8_t res64;
@@ -14714,7 +14716,7 @@ _NEON2SSE_INLINE int8x8_t vneg_s8(int8x8_t a)
     return64(res);
 }
 
-int16x4_t vneg_s16(int16x4_t a); // VNE//d0,d0
+_NEON2SSESTORAGE int16x4_t vneg_s16(int16x4_t a); // VNE//d0,d0
 _NEON2SSE_INLINE int16x4_t vneg_s16(int16x4_t a)
 {
     int16x4_t res64;
@@ -14723,7 +14725,7 @@ _NEON2SSE_INLINE int16x4_t vneg_s16(int16x4_t a)
     return64(res);
 }
 
-int32x2_t vneg_s32(int32x2_t a); // VNE//d0,d0
+_NEON2SSESTORAGE int32x2_t vneg_s32(int32x2_t a); // VNE//d0,d0
 _NEON2SSE_INLINE int32x2_t vneg_s32(int32x2_t a)
 {
     int32x2_t res64;
@@ -14732,18 +14734,18 @@ _NEON2SSE_INLINE int32x2_t vneg_s32(int32x2_t a)
     return64(res);
 }
 
-float32x2_t vneg_f32(float32x2_t a); // VNE//d0,d0
+_NEON2SSESTORAGE float32x2_t vneg_f32(float32x2_t a); // VNE//d0,d0
 _NEON2SSE_INLINE float32x2_t vneg_f32(float32x2_t a) // VNE//d0,d0
 {
     float32x4_t res;
     __m64_128 res64;
-    _NEON2SSE_ALIGN_16 uint32_t c80000000[4] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
+    _NEON2SSE_ALIGN_16 static const uint32_t c80000000[4] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
     res = _mm_xor_ps (_pM128(a), *(__m128*) c80000000); //use low 64 bits
     _M64f(res64, res);
     return res64;
 }
 
-int8x16_t vnegq_s8(int8x16_t a); // VNE//q0,q0
+_NEON2SSESTORAGE int8x16_t vnegq_s8(int8x16_t a); // VNE//q0,q0
 _NEON2SSE_INLINE int8x16_t vnegq_s8(int8x16_t a) // VNE//q0,q0
 {
     __m128i zero;
@@ -14751,7 +14753,7 @@ _NEON2SSE_INLINE int8x16_t vnegq_s8(int8x16_t a) // VNE//q0,q0
     return _mm_sub_epi8 (zero, a);
 } //or _mm_sign_epi8 (a, negative numbers vector)
 
-int16x8_t vnegq_s16(int16x8_t a); // VNE//q0,q0
+_NEON2SSESTORAGE int16x8_t vnegq_s16(int16x8_t a); // VNE//q0,q0
 _NEON2SSE_INLINE int16x8_t vnegq_s16(int16x8_t a) // VNE//q0,q0
 {
     __m128i zero;
@@ -14759,7 +14761,7 @@ _NEON2SSE_INLINE int16x8_t vnegq_s16(int16x8_t a) // VNE//q0,q0
     return _mm_sub_epi16 (zero, a);
 } //or _mm_sign_epi16 (a, negative numbers vector)
 
-int32x4_t vnegq_s32(int32x4_t a); // VNE//q0,q0
+_NEON2SSESTORAGE int32x4_t vnegq_s32(int32x4_t a); // VNE//q0,q0
 _NEON2SSE_INLINE int32x4_t vnegq_s32(int32x4_t a) // VNE//q0,q0
 {
     __m128i zero;
@@ -14767,17 +14769,17 @@ _NEON2SSE_INLINE int32x4_t vnegq_s32(int32x4_t a) // VNE//q0,q0
     return _mm_sub_epi32 (zero, a);
 } //or _mm_sign_epi32 (a, negative numbers vector)
 
-float32x4_t vnegq_f32(float32x4_t a); // VNE//q0,q0
+_NEON2SSESTORAGE float32x4_t vnegq_f32(float32x4_t a); // VNE//q0,q0
 _NEON2SSE_INLINE float32x4_t vnegq_f32(float32x4_t a) // VNE//q0,q0
 {
-    _NEON2SSE_ALIGN_16 uint32_t c80000000[4] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
+    _NEON2SSE_ALIGN_16 static const uint32_t c80000000[4] = {0x80000000, 0x80000000, 0x80000000, 0x80000000};
     return _mm_xor_ps (a, *(__m128*) c80000000);
 }
 
 //************** Saturating Negate: sat(Vd[i] = - Va[i]) **************************
 //***************************************************************************************
 //For signed-integer data types, the negation of the most negative value can't be produced without saturation, while with saturation it is max positive
-int8x8_t vqneg_s8(int8x8_t a); // VQNE//d0,d0
+_NEON2SSESTORAGE int8x8_t vqneg_s8(int8x8_t a); // VQNE//d0,d0
 _NEON2SSE_INLINE int8x8_t vqneg_s8(int8x8_t a)
 {
     int8x8_t res64;
@@ -14786,7 +14788,7 @@ _NEON2SSE_INLINE int8x8_t vqneg_s8(int8x8_t a)
     return64(res);
 }
 
-int16x4_t vqneg_s16(int16x4_t a); // VQNE//d0,d0
+_NEON2SSESTORAGE int16x4_t vqneg_s16(int16x4_t a); // VQNE//d0,d0
 _NEON2SSE_INLINE int16x4_t vqneg_s16(int16x4_t a)
 {
     int16x4_t res64;
@@ -14795,7 +14797,7 @@ _NEON2SSE_INLINE int16x4_t vqneg_s16(int16x4_t a)
     return64(res);
 }
 
-int32x2_t vqneg_s32(int32x2_t a); // VQNE//d0,d0
+_NEON2SSESTORAGE int32x2_t vqneg_s32(int32x2_t a); // VQNE//d0,d0
 _NEON2SSE_INLINE int32x2_t vqneg_s32(int32x2_t a)
 {
     int32x2_t res64;
@@ -14804,7 +14806,7 @@ _NEON2SSE_INLINE int32x2_t vqneg_s32(int32x2_t a)
     return64(res);
 }
 
-int8x16_t vqnegq_s8(int8x16_t a); // VQNE//q0,q0
+_NEON2SSESTORAGE int8x16_t vqnegq_s8(int8x16_t a); // VQNE//q0,q0
 _NEON2SSE_INLINE int8x16_t vqnegq_s8(int8x16_t a) // VQNE//q0,q0
 {
     __m128i zero;
@@ -14812,7 +14814,7 @@ _NEON2SSE_INLINE int8x16_t vqnegq_s8(int8x16_t a) // VQNE//q0,q0
     return _mm_subs_epi8 (zero, a); //saturating substraction
 }
 
-int16x8_t vqnegq_s16(int16x8_t a); // VQNE//q0,q0
+_NEON2SSESTORAGE int16x8_t vqnegq_s16(int16x8_t a); // VQNE//q0,q0
 _NEON2SSE_INLINE int16x8_t vqnegq_s16(int16x8_t a) // VQNE//q0,q0
 {
     __m128i zero;
@@ -14820,7 +14822,7 @@ _NEON2SSE_INLINE int16x8_t vqnegq_s16(int16x8_t a) // VQNE//q0,q0
     return _mm_subs_epi16 (zero, a); //saturating substraction
 }
 
-int32x4_t vqnegq_s32(int32x4_t a); // VQNE//q0,q0
+_NEON2SSESTORAGE int32x4_t vqnegq_s32(int32x4_t a); // VQNE//q0,q0
 _NEON2SSE_INLINE int32x4_t vqnegq_s32(int32x4_t a) // VQNE//q0,q0
 {
     //solution may be not optimal compared with a serial
@@ -14835,7 +14837,7 @@ _NEON2SSE_INLINE int32x4_t vqnegq_s32(int32x4_t a) // VQNE//q0,q0
 //****************** Count leading zeros ********************************
 //**************************************************************************
 //no corresponding vector intrinsics in IA32, need to implement it.  While the implementation is effective for 8 bits, it may be not for 16 and 32 bits
-int8x8_t vclz_s8(int8x8_t a); // VCLZ.I8 d0,d0
+_NEON2SSESTORAGE int8x8_t vclz_s8(int8x8_t a); // VCLZ.I8 d0,d0
 _NEON2SSE_INLINE int8x8_t vclz_s8(int8x8_t a)
 {
     int8x8_t res64;
@@ -14844,7 +14846,7 @@ _NEON2SSE_INLINE int8x8_t vclz_s8(int8x8_t a)
     return64(res);
 }
 
-int16x4_t vclz_s16(int16x4_t a); // VCLZ.I16 d0,d0
+_NEON2SSESTORAGE int16x4_t vclz_s16(int16x4_t a); // VCLZ.I16 d0,d0
 _NEON2SSE_INLINE int16x4_t vclz_s16(int16x4_t a)
 {
     int16x4_t res64;
@@ -14853,7 +14855,7 @@ _NEON2SSE_INLINE int16x4_t vclz_s16(int16x4_t a)
     return64(res);
 }
 
-int32x2_t vclz_s32(int32x2_t a); // VCLZ.I32 d0,d0
+_NEON2SSESTORAGE int32x2_t vclz_s32(int32x2_t a); // VCLZ.I32 d0,d0
 _NEON2SSE_INLINE int32x2_t vclz_s32(int32x2_t a)
 {
     int32x2_t res64;
@@ -14863,22 +14865,22 @@ _NEON2SSE_INLINE int32x2_t vclz_s32(int32x2_t a)
 }
 
 
-uint8x8_t vclz_u8(uint8x8_t a); // VCLZ.I8 d0,d0
+_NEON2SSESTORAGE uint8x8_t vclz_u8(uint8x8_t a); // VCLZ.I8 d0,d0
 #define vclz_u8 vclz_s8
 
-uint16x4_t vclz_u16(uint16x4_t a); // VCLZ.I16 d0,d0
+_NEON2SSESTORAGE uint16x4_t vclz_u16(uint16x4_t a); // VCLZ.I16 d0,d0
 #define vclz_u16 vclz_s16
 
-uint32x2_t vclz_u32(uint32x2_t a); // VCLZ.I32 d0,d0
+_NEON2SSESTORAGE uint32x2_t vclz_u32(uint32x2_t a); // VCLZ.I32 d0,d0
 #define vclz_u32 vclz_s32
 
-int8x16_t vclzq_s8(int8x16_t a); // VCLZ.I8 q0,q0
+_NEON2SSESTORAGE int8x16_t vclzq_s8(int8x16_t a); // VCLZ.I8 q0,q0
 _NEON2SSE_INLINE int8x16_t vclzq_s8(int8x16_t a)
 {
-    _NEON2SSE_ALIGN_16 int8_t mask_CLZ[16] = { /* 0 */ 4,/* 1 */ 3,/* 2 */ 2,/* 3 */ 2,
-                                                    /* 4 */ 1,/* 5 */ 1,/* 6 */ 1,/* 7 */ 1,
-                                                    /* 8 */ 0,/* 9 */ 0,/* a */ 0,/* b */ 0,
-                                                    /* c */ 0,/* d */ 0,/* e */ 0,/* f */ 0                          };
+    _NEON2SSE_ALIGN_16 static const int8_t mask_CLZ[16] = { /* 0 */ 4,/* 1 */ 3,/* 2 */ 2,/* 3 */ 2,
+                                                            /* 4 */ 1,/* 5 */ 1,/* 6 */ 1,/* 7 */ 1,
+                                                            /* 8 */ 0,/* 9 */ 0,/* a */ 0,/* b */ 0,
+                                                            /* c */ 0,/* d */ 0,/* e */ 0,/* f */ 0                          };
     __m128i maskLOW, c4, lowclz, mask, hiclz;
     maskLOW = _mm_set1_epi8(0x0f); //low 4 bits, don't need masking low to avoid zero if MSB is set - it happens automatically
     c4 = _mm_set1_epi8(4);
@@ -14891,12 +14893,12 @@ _NEON2SSE_INLINE int8x16_t vclzq_s8(int8x16_t a)
     return _mm_add_epi8(lowclz, hiclz);
 }
 
-int16x8_t vclzq_s16(int16x8_t a); // VCLZ.I16 q0,q0
+_NEON2SSESTORAGE int16x8_t vclzq_s16(int16x8_t a); // VCLZ.I16 q0,q0
 _NEON2SSE_INLINE int16x8_t vclzq_s16(int16x8_t a)
 {
     __m128i c7, res8x16, res8x16_swap;
-    _NEON2SSE_ALIGN_16 int8_t mask8_sab[16] = { 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14};
-    _NEON2SSE_ALIGN_16 uint16_t mask8bit[8] = {0x00ff, 0x00ff, 0x00ff, 0x00ff,0x00ff, 0x00ff, 0x00ff, 0x00ff};
+    _NEON2SSE_ALIGN_16 static const int8_t mask8_sab[16] = { 1, 0, 3, 2, 5, 4, 7, 6, 9, 8, 11, 10, 13, 12, 15, 14};
+    _NEON2SSE_ALIGN_16 static const uint16_t mask8bit[8] = {0x00ff, 0x00ff, 0x00ff, 0x00ff,0x00ff, 0x00ff, 0x00ff, 0x00ff};
     c7 = _mm_srli_epi16(*(__m128i*)mask8bit, 5); //7
     res8x16 = vclzq_s8(a);
     res8x16_swap = _mm_shuffle_epi8 (res8x16, *(__m128i*) mask8_sab); //horisontal pairs swap
@@ -14907,7 +14909,7 @@ _NEON2SSE_INLINE int16x8_t vclzq_s16(int16x8_t a)
     return _mm_add_epi16(res8x16_swap, res8x16);
 }
 
-int32x4_t vclzq_s32(int32x4_t a); // VCLZ.I32 q0,q0
+_NEON2SSESTORAGE int32x4_t vclzq_s32(int32x4_t a); // VCLZ.I32 q0,q0
 _NEON2SSE_INLINE int32x4_t vclzq_s32(int32x4_t a)
 {
     __m128i c55555555, c33333333, c0f0f0f0f, c3f, c32, tmp, tmp1, res;
@@ -14951,13 +14953,13 @@ _NEON2SSE_INLINE int32x4_t vclzq_s32(int32x4_t a)
     return _mm_sub_epi32(c32, res); //res[i] = 32 - atmp[i];
 }
 
-uint8x16_t vclzq_u8(uint8x16_t a); // VCLZ.I8 q0,q0
+_NEON2SSESTORAGE uint8x16_t vclzq_u8(uint8x16_t a); // VCLZ.I8 q0,q0
 #define vclzq_u8 vclzq_s8
 
-uint16x8_t vclzq_u16(uint16x8_t a); // VCLZ.I16 q0,q0
+_NEON2SSESTORAGE uint16x8_t vclzq_u16(uint16x8_t a); // VCLZ.I16 q0,q0
 #define vclzq_u16 vclzq_s16
 
-uint32x4_t vclzq_u32(uint32x4_t a); // VCLZ.I32 q0,q0
+_NEON2SSESTORAGE uint32x4_t vclzq_u32(uint32x4_t a); // VCLZ.I32 q0,q0
 #define vclzq_u32 vclzq_s32
 
 //************** Count leading sign bits **************************
@@ -14966,7 +14968,7 @@ uint32x4_t vclzq_u32(uint32x4_t a); // VCLZ.I32 q0,q0
 // the topmost bit, that are the same as the topmost bit, in each element in a vector
 //No corresponding vector intrinsics in IA32, need to implement it.
 //While the implementation is effective for 8 bits, it may be not for 16 and 32 bits
-int8x8_t vcls_s8(int8x8_t a); // VCLS.S8 d0,d0
+_NEON2SSESTORAGE int8x8_t vcls_s8(int8x8_t a); // VCLS.S8 d0,d0
 _NEON2SSE_INLINE int8x8_t vcls_s8(int8x8_t a)
 {
     int8x8_t res64;
@@ -14975,7 +14977,7 @@ _NEON2SSE_INLINE int8x8_t vcls_s8(int8x8_t a)
     return64(res);
 }
 
-int16x4_t vcls_s16(int16x4_t a); // VCLS.S16 d0,d0
+_NEON2SSESTORAGE int16x4_t vcls_s16(int16x4_t a); // VCLS.S16 d0,d0
 _NEON2SSE_INLINE int16x4_t vcls_s16(int16x4_t a)
 {
     int16x4_t res64;
@@ -14984,7 +14986,7 @@ _NEON2SSE_INLINE int16x4_t vcls_s16(int16x4_t a)
     return64(res);
 }
 
-int32x2_t vcls_s32(int32x2_t a); // VCLS.S32 d0,d0
+_NEON2SSESTORAGE int32x2_t vcls_s32(int32x2_t a); // VCLS.S32 d0,d0
 _NEON2SSE_INLINE int32x2_t vcls_s32(int32x2_t a)
 {
     int32x2_t res64;
@@ -14993,7 +14995,7 @@ _NEON2SSE_INLINE int32x2_t vcls_s32(int32x2_t a)
     return64(res);
 }
 
-int8x16_t vclsq_s8(int8x16_t a); // VCLS.S8 q0,q0
+_NEON2SSESTORAGE int8x16_t vclsq_s8(int8x16_t a); // VCLS.S8 q0,q0
 _NEON2SSE_INLINE int8x16_t vclsq_s8(int8x16_t a)
 {
     __m128i cff, c80, c1, a_mask, a_neg, a_pos, a_comb;
@@ -15010,7 +15012,7 @@ _NEON2SSE_INLINE int8x16_t vclsq_s8(int8x16_t a)
     return _mm_sub_epi8(a_comb, c1);
 }
 
-int16x8_t vclsq_s16(int16x8_t a); // VCLS.S16 q0,q0
+_NEON2SSESTORAGE int16x8_t vclsq_s16(int16x8_t a); // VCLS.S16 q0,q0
 _NEON2SSE_INLINE int16x8_t vclsq_s16(int16x8_t a)
 {
     __m128i cffff, c8000, c1, a_mask, a_neg, a_pos, a_comb;
@@ -15027,7 +15029,7 @@ _NEON2SSE_INLINE int16x8_t vclsq_s16(int16x8_t a)
     return _mm_sub_epi16(a_comb, c1);
 }
 
-int32x4_t vclsq_s32(int32x4_t a); // VCLS.S32 q0,q0
+_NEON2SSESTORAGE int32x4_t vclsq_s32(int32x4_t a); // VCLS.S32 q0,q0
 _NEON2SSE_INLINE int32x4_t vclsq_s32(int32x4_t a)
 {
     __m128i cffffffff, c80000000, c1, a_mask, a_neg, a_pos, a_comb;
@@ -15049,7 +15051,7 @@ _NEON2SSE_INLINE int32x4_t vclsq_s32(int32x4_t a)
 //No corresponding SIMD solution. One option is to get a elements, convert it to 32 bits and then use SSE4.2  _mm_popcnt__u32 (unsigned int v) for each element
 //another option is to do the following algorithm:
 
-uint8x8_t vcnt_u8(uint8x8_t a); // VCNT.8 d0,d0
+_NEON2SSESTORAGE uint8x8_t vcnt_u8(uint8x8_t a); // VCNT.8 d0,d0
 _NEON2SSE_INLINE uint8x8_t vcnt_u8(uint8x8_t a)
 {
     uint8x8_t res64;
@@ -15058,19 +15060,19 @@ _NEON2SSE_INLINE uint8x8_t vcnt_u8(uint8x8_t a)
     return64(res);
 }
 
-int8x8_t vcnt_s8(int8x8_t a); // VCNT.8 d0,d0
+_NEON2SSESTORAGE int8x8_t vcnt_s8(int8x8_t a); // VCNT.8 d0,d0
 #define vcnt_s8 vcnt_u8
 
-poly8x8_t vcnt_p8(poly8x8_t a); // VCNT.8 d0,d0
+_NEON2SSESTORAGE poly8x8_t vcnt_p8(poly8x8_t a); // VCNT.8 d0,d0
 #define vcnt_p8 vcnt_u8
 
-uint8x16_t vcntq_u8(uint8x16_t a); // VCNT.8 q0,q0
+_NEON2SSESTORAGE uint8x16_t vcntq_u8(uint8x16_t a); // VCNT.8 q0,q0
 _NEON2SSE_INLINE uint8x16_t vcntq_u8(uint8x16_t a)
 {
-    _NEON2SSE_ALIGN_16 int8_t mask_POPCOUNT[16] = { /* 0 */ 0,/* 1 */ 1,/* 2 */ 1,/* 3 */ 2,
-                                                        /* 4 */ 1,/* 5 */ 2,/* 6 */ 2,/* 7 */ 3,
-                                                        /* 8 */ 1,/* 9 */ 2,/* a */ 2,/* b */ 3,
-                                                        /* c */ 2,/* d */ 3,/* e */ 3,/* f */ 4                                   };
+    _NEON2SSE_ALIGN_16 static const int8_t mask_POPCOUNT[16] = { /* 0 */ 0,/* 1 */ 1,/* 2 */ 1,/* 3 */ 2,
+                                                                 /* 4 */ 1,/* 5 */ 2,/* 6 */ 2,/* 7 */ 3,
+                                                                 /* 8 */ 1,/* 9 */ 2,/* a */ 2,/* b */ 3,
+                                                                 /* c */ 2,/* d */ 3,/* e */ 3,/* f */ 4};
     __m128i maskLOW, mask, lowpopcnt, hipopcnt;
     maskLOW = _mm_set1_epi8(0x0f); //low 4 bits, need masking to avoid zero if MSB is set
     mask = _mm_and_si128(a, maskLOW);
@@ -15081,10 +15083,10 @@ _NEON2SSE_INLINE uint8x16_t vcntq_u8(uint8x16_t a)
     return _mm_add_epi8(lowpopcnt, hipopcnt);
 }
 
-int8x16_t vcntq_s8(int8x16_t a); // VCNT.8 q0,q0
+_NEON2SSESTORAGE int8x16_t vcntq_s8(int8x16_t a); // VCNT.8 q0,q0
 #define vcntq_s8 vcntq_u8
 
-poly8x16_t vcntq_p8(poly8x16_t a); // VCNT.8 q0,q0
+_NEON2SSESTORAGE poly8x16_t vcntq_p8(poly8x16_t a); // VCNT.8 q0,q0
 #define vcntq_p8 vcntq_u8
 
 //**************************************************************************************
@@ -15092,7 +15094,7 @@ poly8x16_t vcntq_p8(poly8x16_t a); // VCNT.8 q0,q0
 //**************************************************************************************
 //************************** Bitwise not ***********************************
 //several Bitwise not implementations possible for SIMD. Eg "xor" with all ones, but the following one gives good performance
-int8x8_t vmvn_s8(int8x8_t a); // VMVN d0,d0
+_NEON2SSESTORAGE int8x8_t vmvn_s8(int8x8_t a); // VMVN d0,d0
 _NEON2SSE_INLINE int8x8_t vmvn_s8(int8x8_t a)
 {
     int8x8_t res64;
@@ -15101,7 +15103,7 @@ _NEON2SSE_INLINE int8x8_t vmvn_s8(int8x8_t a)
     return64(res);
 }
 
-int16x4_t vmvn_s16(int16x4_t a); // VMVN d0,d0
+_NEON2SSESTORAGE int16x4_t vmvn_s16(int16x4_t a); // VMVN d0,d0
 _NEON2SSE_INLINE int16x4_t vmvn_s16(int16x4_t a)
 {
     int16x4_t res64;
@@ -15110,7 +15112,7 @@ _NEON2SSE_INLINE int16x4_t vmvn_s16(int16x4_t a)
     return64(res);
 }
 
-int32x2_t vmvn_s32(int32x2_t a); // VMVN d0,d0
+_NEON2SSESTORAGE int32x2_t vmvn_s32(int32x2_t a); // VMVN d0,d0
 _NEON2SSE_INLINE int32x2_t vmvn_s32(int32x2_t a)
 {
     int32x2_t res64;
@@ -15119,19 +15121,19 @@ _NEON2SSE_INLINE int32x2_t vmvn_s32(int32x2_t a)
     return64(res);
 }
 
-uint8x8_t vmvn_u8(uint8x8_t a); // VMVN d0,d0
+_NEON2SSESTORAGE uint8x8_t vmvn_u8(uint8x8_t a); // VMVN d0,d0
 #define vmvn_u8 vmvn_s8
 
-uint16x4_t vmvn_u16(uint16x4_t a); // VMVN d0,d0
+_NEON2SSESTORAGE uint16x4_t vmvn_u16(uint16x4_t a); // VMVN d0,d0
 #define vmvn_u16 vmvn_s16
 
-uint32x2_t vmvn_u32(uint32x2_t a); // VMVN d0,d0
+_NEON2SSESTORAGE uint32x2_t vmvn_u32(uint32x2_t a); // VMVN d0,d0
 #define vmvn_u32 vmvn_s32
 
-poly8x8_t vmvn_p8(poly8x8_t a); // VMVN d0,d0
+_NEON2SSESTORAGE poly8x8_t vmvn_p8(poly8x8_t a); // VMVN d0,d0
 #define vmvn_p8 vmvn_u8
 
-int8x16_t vmvnq_s8(int8x16_t a); // VMVN q0,q0
+_NEON2SSESTORAGE int8x16_t vmvnq_s8(int8x16_t a); // VMVN q0,q0
 _NEON2SSE_INLINE int8x16_t vmvnq_s8(int8x16_t a) // VMVN q0,q0
 {
     __m128i c1;
@@ -15139,7 +15141,7 @@ _NEON2SSE_INLINE int8x16_t vmvnq_s8(int8x16_t a) // VMVN q0,q0
     return _mm_andnot_si128 (a, c1);
 }
 
-int16x8_t vmvnq_s16(int16x8_t a); // VMVN q0,q0
+_NEON2SSESTORAGE int16x8_t vmvnq_s16(int16x8_t a); // VMVN q0,q0
 _NEON2SSE_INLINE int16x8_t vmvnq_s16(int16x8_t a) // VMVN q0,q0
 {
     __m128i c1;
@@ -15147,7 +15149,7 @@ _NEON2SSE_INLINE int16x8_t vmvnq_s16(int16x8_t a) // VMVN q0,q0
     return _mm_andnot_si128 (a, c1);
 }
 
-int32x4_t vmvnq_s32(int32x4_t a); // VMVN q0,q0
+_NEON2SSESTORAGE int32x4_t vmvnq_s32(int32x4_t a); // VMVN q0,q0
 _NEON2SSE_INLINE int32x4_t vmvnq_s32(int32x4_t a) // VMVN q0,q0
 {
     __m128i c1;
@@ -15155,35 +15157,35 @@ _NEON2SSE_INLINE int32x4_t vmvnq_s32(int32x4_t a) // VMVN q0,q0
     return _mm_andnot_si128 (a, c1);
 }
 
-uint8x16_t vmvnq_u8(uint8x16_t a); // VMVN q0,q0
+_NEON2SSESTORAGE uint8x16_t vmvnq_u8(uint8x16_t a); // VMVN q0,q0
 #define vmvnq_u8 vmvnq_s8
 
-uint16x8_t vmvnq_u16(uint16x8_t a); // VMVN q0,q0
+_NEON2SSESTORAGE uint16x8_t vmvnq_u16(uint16x8_t a); // VMVN q0,q0
 #define vmvnq_u16 vmvnq_s16
 
-uint32x4_t vmvnq_u32(uint32x4_t a); // VMVN q0,q0
+_NEON2SSESTORAGE uint32x4_t vmvnq_u32(uint32x4_t a); // VMVN q0,q0
 #define vmvnq_u32 vmvnq_s32
 
-poly8x16_t vmvnq_p8(poly8x16_t a); // VMVN q0,q0
+_NEON2SSESTORAGE poly8x16_t vmvnq_p8(poly8x16_t a); // VMVN q0,q0
 #define vmvnq_p8 vmvnq_u8
 
 //****************** Bitwise and ***********************
 //******************************************************
-int8x8_t vand_s8(int8x8_t a, int8x8_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vand_s8(int8x8_t a, int8x8_t b); // VAND d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vand_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
     return64(_mm_and_si128(_pM128i(a),_pM128i(b)));
 }
 
-int16x4_t vand_s16(int16x4_t a, int16x4_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vand_s16(int16x4_t a, int16x4_t b); // VAND d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vand_s16(int16x4_t a, int16x4_t b)
 {
     int16x4_t res64;
     return64(_mm_and_si128(_pM128i(a),_pM128i(b)));
 }
 
-int32x2_t vand_s32(int32x2_t a, int32x2_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vand_s32(int32x2_t a, int32x2_t b); // VAND d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vand_s32(int32x2_t a, int32x2_t b)
 {
     int32x2_t res64;
@@ -15191,7 +15193,7 @@ _NEON2SSE_INLINE int32x2_t vand_s32(int32x2_t a, int32x2_t b)
 }
 
 
-int64x1_t vand_s64(int64x1_t a,  int64x1_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vand_s64(int64x1_t a,  int64x1_t b); // VAND d0,d0,d0
 _NEON2SSE_INLINE int64x1_t vand_s64(int64x1_t a,  int64x1_t b)
 {
     int64x1_t res;
@@ -15199,46 +15201,46 @@ _NEON2SSE_INLINE int64x1_t vand_s64(int64x1_t a,  int64x1_t b)
     return res;
 }
 
-uint8x8_t vand_u8(uint8x8_t a, uint8x8_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vand_u8(uint8x8_t a, uint8x8_t b); // VAND d0,d0,d0
 #define vand_u8 vand_s8
 
-uint16x4_t vand_u16(uint16x4_t a, uint16x4_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vand_u16(uint16x4_t a, uint16x4_t b); // VAND d0,d0,d0
 #define vand_u16 vand_s16
 
-uint32x2_t vand_u32(uint32x2_t a, uint32x2_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vand_u32(uint32x2_t a, uint32x2_t b); // VAND d0,d0,d0
 #define vand_u32 vand_s32
 
-uint64x1_t vand_u64(uint64x1_t a,  uint64x1_t b); // VAND d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vand_u64(uint64x1_t a,  uint64x1_t b); // VAND d0,d0,d0
 #define vand_u64 vand_s64
 
 
-int8x16_t   vandq_s8(int8x16_t a, int8x16_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE int8x16_t   vandq_s8(int8x16_t a, int8x16_t b); // VAND q0,q0,q0
 #define vandq_s8 _mm_and_si128
 
-int16x8_t   vandq_s16(int16x8_t a, int16x8_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE int16x8_t   vandq_s16(int16x8_t a, int16x8_t b); // VAND q0,q0,q0
 #define vandq_s16 _mm_and_si128
 
-int32x4_t   vandq_s32(int32x4_t a, int32x4_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE int32x4_t   vandq_s32(int32x4_t a, int32x4_t b); // VAND q0,q0,q0
 #define vandq_s32 _mm_and_si128
 
-int64x2_t   vandq_s64(int64x2_t a, int64x2_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE int64x2_t   vandq_s64(int64x2_t a, int64x2_t b); // VAND q0,q0,q0
 #define vandq_s64 _mm_and_si128
 
-uint8x16_t   vandq_u8(uint8x16_t a, uint8x16_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t   vandq_u8(uint8x16_t a, uint8x16_t b); // VAND q0,q0,q0
 #define vandq_u8 _mm_and_si128
 
-uint16x8_t   vandq_u16(uint16x8_t a, uint16x8_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   vandq_u16(uint16x8_t a, uint16x8_t b); // VAND q0,q0,q0
 #define vandq_u16 _mm_and_si128
 
-uint32x4_t   vandq_u32(uint32x4_t a, uint32x4_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t   vandq_u32(uint32x4_t a, uint32x4_t b); // VAND q0,q0,q0
 #define vandq_u32 _mm_and_si128
 
-uint64x2_t   vandq_u64(uint64x2_t a, uint64x2_t b); // VAND q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t   vandq_u64(uint64x2_t a, uint64x2_t b); // VAND q0,q0,q0
 #define vandq_u64 _mm_and_si128
 
 //******************** Bitwise or *********************************
 //******************************************************************
-int8x8_t vorr_s8(int8x8_t a, int8x8_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vorr_s8(int8x8_t a, int8x8_t b); // VORR d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vorr_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
@@ -15246,7 +15248,7 @@ _NEON2SSE_INLINE int8x8_t vorr_s8(int8x8_t a, int8x8_t b)
 }
 
 
-int16x4_t vorr_s16(int16x4_t a, int16x4_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vorr_s16(int16x4_t a, int16x4_t b); // VORR d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vorr_s16(int16x4_t a, int16x4_t b)
 {
     int16x4_t res64;
@@ -15254,7 +15256,7 @@ _NEON2SSE_INLINE int16x4_t vorr_s16(int16x4_t a, int16x4_t b)
 }
 
 
-int32x2_t vorr_s32(int32x2_t a, int32x2_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vorr_s32(int32x2_t a, int32x2_t b); // VORR d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vorr_s32(int32x2_t a, int32x2_t b)
 {
     int32x2_t res64;
@@ -15262,7 +15264,7 @@ _NEON2SSE_INLINE int32x2_t vorr_s32(int32x2_t a, int32x2_t b)
 }
 
 
-int64x1_t vorr_s64(int64x1_t a,  int64x1_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vorr_s64(int64x1_t a,  int64x1_t b); // VORR d0,d0,d0
 _NEON2SSE_INLINE int64x1_t vorr_s64(int64x1_t a,  int64x1_t b)
 {
     int64x1_t res;
@@ -15270,58 +15272,58 @@ _NEON2SSE_INLINE int64x1_t vorr_s64(int64x1_t a,  int64x1_t b)
     return res;
 }
 
-uint8x8_t vorr_u8(uint8x8_t a, uint8x8_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vorr_u8(uint8x8_t a, uint8x8_t b); // VORR d0,d0,d0
 #define vorr_u8 vorr_s8
 
-uint16x4_t vorr_u16(uint16x4_t a, uint16x4_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vorr_u16(uint16x4_t a, uint16x4_t b); // VORR d0,d0,d0
 #define vorr_u16 vorr_s16
 
-uint32x2_t vorr_u32(uint32x2_t a, uint32x2_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vorr_u32(uint32x2_t a, uint32x2_t b); // VORR d0,d0,d0
 #define vorr_u32 vorr_s32
 
-uint64x1_t vorr_u64(uint64x1_t a,  uint64x1_t b); // VORR d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vorr_u64(uint64x1_t a,  uint64x1_t b); // VORR d0,d0,d0
 #define vorr_u64 vorr_s64
 
-int8x16_t   vorrq_s8(int8x16_t a, int8x16_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE int8x16_t   vorrq_s8(int8x16_t a, int8x16_t b); // VORR q0,q0,q0
 #define vorrq_s8 _mm_or_si128
 
-int16x8_t   vorrq_s16(int16x8_t a, int16x8_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE int16x8_t   vorrq_s16(int16x8_t a, int16x8_t b); // VORR q0,q0,q0
 #define vorrq_s16 _mm_or_si128
 
-int32x4_t   vorrq_s32(int32x4_t a, int32x4_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE int32x4_t   vorrq_s32(int32x4_t a, int32x4_t b); // VORR q0,q0,q0
 #define vorrq_s32 _mm_or_si128
 
-int64x2_t   vorrq_s64(int64x2_t a, int64x2_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE int64x2_t   vorrq_s64(int64x2_t a, int64x2_t b); // VORR q0,q0,q0
 #define vorrq_s64 _mm_or_si128
 
-uint8x16_t   vorrq_u8(uint8x16_t a, uint8x16_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t   vorrq_u8(uint8x16_t a, uint8x16_t b); // VORR q0,q0,q0
 #define vorrq_u8 _mm_or_si128
 
-uint16x8_t   vorrq_u16(uint16x8_t a, uint16x8_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   vorrq_u16(uint16x8_t a, uint16x8_t b); // VORR q0,q0,q0
 #define vorrq_u16 _mm_or_si128
 
-uint32x4_t   vorrq_u32(uint32x4_t a, uint32x4_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t   vorrq_u32(uint32x4_t a, uint32x4_t b); // VORR q0,q0,q0
 #define vorrq_u32 _mm_or_si128
 
-uint64x2_t   vorrq_u64(uint64x2_t a, uint64x2_t b); // VORR q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t   vorrq_u64(uint64x2_t a, uint64x2_t b); // VORR q0,q0,q0
 #define vorrq_u64 _mm_or_si128
 
 //************* Bitwise exclusive or (EOR or XOR) ******************
 //*******************************************************************
-int8x8_t veor_s8(int8x8_t a, int8x8_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE int8x8_t veor_s8(int8x8_t a, int8x8_t b); // VEOR d0,d0,d0
 _NEON2SSE_INLINE int8x8_t veor_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
     return64(_mm_xor_si128(_pM128i(a),_pM128i(b)));
 }
 
-int16x4_t veor_s16(int16x4_t a, int16x4_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE int16x4_t veor_s16(int16x4_t a, int16x4_t b); // VEOR d0,d0,d0
 #define veor_s16 veor_s8
 
-int32x2_t veor_s32(int32x2_t a, int32x2_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE int32x2_t veor_s32(int32x2_t a, int32x2_t b); // VEOR d0,d0,d0
 #define veor_s32 veor_s8
 
-int64x1_t veor_s64(int64x1_t a,  int64x1_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE int64x1_t veor_s64(int64x1_t a,  int64x1_t b); // VEOR d0,d0,d0
 _NEON2SSE_INLINE int64x1_t veor_s64(int64x1_t a,  int64x1_t b)
 {
     int64x1_t res;
@@ -15329,59 +15331,59 @@ _NEON2SSE_INLINE int64x1_t veor_s64(int64x1_t a,  int64x1_t b)
     return res;
 }
 
-uint8x8_t veor_u8(uint8x8_t a, uint8x8_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t veor_u8(uint8x8_t a, uint8x8_t b); // VEOR d0,d0,d0
 #define veor_u8 veor_s8
 
-uint16x4_t veor_u16(uint16x4_t a, uint16x4_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t veor_u16(uint16x4_t a, uint16x4_t b); // VEOR d0,d0,d0
 #define veor_u16 veor_s16
 
-uint32x2_t veor_u32(uint32x2_t a, uint32x2_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t veor_u32(uint32x2_t a, uint32x2_t b); // VEOR d0,d0,d0
 #define veor_u32 veor_s32
 
-uint64x1_t veor_u64(uint64x1_t a,  uint64x1_t b); // VEOR d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t veor_u64(uint64x1_t a,  uint64x1_t b); // VEOR d0,d0,d0
 #define veor_u64 veor_s64
 
-int8x16_t   veorq_s8(int8x16_t a, int8x16_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE int8x16_t   veorq_s8(int8x16_t a, int8x16_t b); // VEOR q0,q0,q0
 #define veorq_s8 _mm_xor_si128
 
-int16x8_t   veorq_s16(int16x8_t a, int16x8_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE int16x8_t   veorq_s16(int16x8_t a, int16x8_t b); // VEOR q0,q0,q0
 #define veorq_s16 _mm_xor_si128
 
-int32x4_t   veorq_s32(int32x4_t a, int32x4_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE int32x4_t   veorq_s32(int32x4_t a, int32x4_t b); // VEOR q0,q0,q0
 #define veorq_s32 _mm_xor_si128
 
-int64x2_t   veorq_s64(int64x2_t a, int64x2_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE int64x2_t   veorq_s64(int64x2_t a, int64x2_t b); // VEOR q0,q0,q0
 #define veorq_s64 _mm_xor_si128
 
-uint8x16_t   veorq_u8(uint8x16_t a, uint8x16_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t   veorq_u8(uint8x16_t a, uint8x16_t b); // VEOR q0,q0,q0
 #define veorq_u8 _mm_xor_si128
 
-uint16x8_t   veorq_u16(uint16x8_t a, uint16x8_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   veorq_u16(uint16x8_t a, uint16x8_t b); // VEOR q0,q0,q0
 #define veorq_u16 _mm_xor_si128
 
-uint32x4_t   veorq_u32(uint32x4_t a, uint32x4_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t   veorq_u32(uint32x4_t a, uint32x4_t b); // VEOR q0,q0,q0
 #define veorq_u32 _mm_xor_si128
 
-uint64x2_t   veorq_u64(uint64x2_t a, uint64x2_t b); // VEOR q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t   veorq_u64(uint64x2_t a, uint64x2_t b); // VEOR q0,q0,q0
 #define veorq_u64 _mm_xor_si128
 
 //********************** Bit Clear **********************************
 //*******************************************************************
 //Logical AND complement (AND negation or AND NOT)
-int8x8_t   vbic_s8(int8x8_t a, int8x8_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE int8x8_t   vbic_s8(int8x8_t a, int8x8_t b); // VBIC d0,d0,d0
 _NEON2SSE_INLINE int8x8_t   vbic_s8(int8x8_t a, int8x8_t b)
 {
     int8x8_t res64;
     return64(_mm_andnot_si128(_pM128i(b),_pM128i(a))); //notice the arguments "swap"
 }
 
-int16x4_t   vbic_s16(int16x4_t a, int16x4_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE int16x4_t   vbic_s16(int16x4_t a, int16x4_t b); // VBIC d0,d0,d0
 #define vbic_s16 vbic_s8
 
-int32x2_t   vbic_s32(int32x2_t a, int32x2_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE int32x2_t   vbic_s32(int32x2_t a, int32x2_t b); // VBIC d0,d0,d0
 #define vbic_s32 vbic_s8
 
-int64x1_t   vbic_s64(int64x1_t a, int64x1_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE int64x1_t   vbic_s64(int64x1_t a, int64x1_t b); // VBIC d0,d0,d0
 _NEON2SSE_INLINE int64x1_t   vbic_s64(int64x1_t a, int64x1_t b)
 {
     int64x1_t res;
@@ -15389,46 +15391,46 @@ _NEON2SSE_INLINE int64x1_t   vbic_s64(int64x1_t a, int64x1_t b)
     return res;
 }
 
-uint8x8_t   vbic_u8(uint8x8_t a, uint8x8_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t   vbic_u8(uint8x8_t a, uint8x8_t b); // VBIC d0,d0,d0
 #define vbic_u8 vbic_s8
 
-uint16x4_t   vbic_u16(uint16x4_t a, uint16x4_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t   vbic_u16(uint16x4_t a, uint16x4_t b); // VBIC d0,d0,d0
 #define vbic_u16 vbic_s16
 
-uint32x2_t   vbic_u32(uint32x2_t a, uint32x2_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t   vbic_u32(uint32x2_t a, uint32x2_t b); // VBIC d0,d0,d0
 #define vbic_u32 vbic_s32
 
-uint64x1_t   vbic_u64(uint64x1_t a, uint64x1_t b); // VBIC d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t   vbic_u64(uint64x1_t a, uint64x1_t b); // VBIC d0,d0,d0
 #define vbic_u64 vbic_s64
 
-int8x16_t   vbicq_s8(int8x16_t a, int8x16_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE int8x16_t   vbicq_s8(int8x16_t a, int8x16_t b); // VBIC q0,q0,q0
 #define vbicq_s8(a,b) _mm_andnot_si128 (b,a) //notice arguments "swap"
 
-int16x8_t   vbicq_s16(int16x8_t a, int16x8_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE int16x8_t   vbicq_s16(int16x8_t a, int16x8_t b); // VBIC q0,q0,q0
 #define vbicq_s16(a,b) _mm_andnot_si128 (b,a) //notice arguments "swap"
 
-int32x4_t   vbicq_s32(int32x4_t a, int32x4_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE int32x4_t   vbicq_s32(int32x4_t a, int32x4_t b); // VBIC q0,q0,q0
 #define vbicq_s32(a,b) _mm_andnot_si128 (b,a) //notice arguments "swap"
 
-int64x2_t   vbicq_s64(int64x2_t a, int64x2_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE int64x2_t   vbicq_s64(int64x2_t a, int64x2_t b); // VBIC q0,q0,q0
 #define vbicq_s64(a,b) _mm_andnot_si128 (b,a) //notice arguments "swap"
 
-uint8x16_t   vbicq_u8(uint8x16_t a, uint8x16_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t   vbicq_u8(uint8x16_t a, uint8x16_t b); // VBIC q0,q0,q0
 #define vbicq_u8(a,b) _mm_andnot_si128 (b,a) //notice arguments "swap"
 
-uint16x8_t   vbicq_u16(uint16x8_t a, uint16x8_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t   vbicq_u16(uint16x8_t a, uint16x8_t b); // VBIC q0,q0,q0
 #define vbicq_u16(a,b) _mm_andnot_si128 (b,a) //notice arguments "swap"
 
-uint32x4_t   vbicq_u32(uint32x4_t a, uint32x4_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t   vbicq_u32(uint32x4_t a, uint32x4_t b); // VBIC q0,q0,q0
 #define vbicq_u32(a,b) _mm_andnot_si128 (b,a) //notice arguments "swap"
 
-uint64x2_t   vbicq_u64(uint64x2_t a, uint64x2_t b); // VBIC q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t   vbicq_u64(uint64x2_t a, uint64x2_t b); // VBIC q0,q0,q0
 #define vbicq_u64(a,b) _mm_andnot_si128 (b,a) //notice arguments "swap"
 
 //**************** Bitwise OR complement ********************************
 //**************************************** ********************************
 //no exact IA 32 match, need to implement it as following
-int8x8_t vorn_s8(int8x8_t a,  int8x8_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vorn_s8(int8x8_t a,  int8x8_t b); // VORN d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vorn_s8(int8x8_t a,  int8x8_t b)
 {
     int8x8_t res64;
@@ -15436,7 +15438,7 @@ _NEON2SSE_INLINE int8x8_t vorn_s8(int8x8_t a,  int8x8_t b)
 }
 
 
-int16x4_t vorn_s16(int16x4_t a,  int16x4_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vorn_s16(int16x4_t a,  int16x4_t b); // VORN d0,d0,d0
 _NEON2SSE_INLINE int16x4_t vorn_s16(int16x4_t a,  int16x4_t b)
 {
     int16x4_t res64;
@@ -15444,7 +15446,7 @@ _NEON2SSE_INLINE int16x4_t vorn_s16(int16x4_t a,  int16x4_t b)
 }
 
 
-int32x2_t vorn_s32(int32x2_t a,  int32x2_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vorn_s32(int32x2_t a,  int32x2_t b); // VORN d0,d0,d0
 _NEON2SSE_INLINE int32x2_t vorn_s32(int32x2_t a,  int32x2_t b)
 {
     int32x2_t res64;
@@ -15452,7 +15454,7 @@ _NEON2SSE_INLINE int32x2_t vorn_s32(int32x2_t a,  int32x2_t b)
 }
 
 
-int64x1_t vorn_s64(int64x1_t a, int64x1_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vorn_s64(int64x1_t a, int64x1_t b); // VORN d0,d0,d0
 _NEON2SSE_INLINE int64x1_t vorn_s64(int64x1_t a, int64x1_t b)
 {
     int64x1_t res;
@@ -15460,21 +15462,21 @@ _NEON2SSE_INLINE int64x1_t vorn_s64(int64x1_t a, int64x1_t b)
     return res;
 }
 
-uint8x8_t vorn_u8(uint8x8_t a,  uint8x8_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vorn_u8(uint8x8_t a,  uint8x8_t b); // VORN d0,d0,d0
 #define vorn_u8 vorn_s8
 
 
-uint16x4_t vorn_u16(uint16x4_t a,  uint16x4_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vorn_u16(uint16x4_t a,  uint16x4_t b); // VORN d0,d0,d0
 #define vorn_u16 vorn_s16
 
-uint32x2_t vorn_u32(uint32x2_t a,  uint32x2_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vorn_u32(uint32x2_t a,  uint32x2_t b); // VORN d0,d0,d0
 #define vorn_u32 vorn_s32
 
-uint64x1_t vorn_u64(uint64x1_t a, uint64x1_t b); // VORN d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vorn_u64(uint64x1_t a, uint64x1_t b); // VORN d0,d0,d0
 #define vorn_u64 vorn_s64
 
 
-int8x16_t vornq_s8(int8x16_t a, int8x16_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vornq_s8(int8x16_t a, int8x16_t b); // VORN q0,q0,q0
 _NEON2SSE_INLINE int8x16_t vornq_s8(int8x16_t a, int8x16_t b) // VORN q0,q0,q0
 {
     __m128i b1;
@@ -15482,7 +15484,7 @@ _NEON2SSE_INLINE int8x16_t vornq_s8(int8x16_t a, int8x16_t b) // VORN q0,q0,q0
     return _mm_or_si128 (a, b1);
 }
 
-int16x8_t vornq_s16(int16x8_t a, int16x8_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vornq_s16(int16x8_t a, int16x8_t b); // VORN q0,q0,q0
 _NEON2SSE_INLINE int16x8_t vornq_s16(int16x8_t a, int16x8_t b) // VORN q0,q0,q0
 {
     __m128i b1;
@@ -15490,7 +15492,7 @@ _NEON2SSE_INLINE int16x8_t vornq_s16(int16x8_t a, int16x8_t b) // VORN q0,q0,q0
     return _mm_or_si128 (a, b1);
 }
 
-int32x4_t vornq_s32(int32x4_t a, int32x4_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vornq_s32(int32x4_t a, int32x4_t b); // VORN q0,q0,q0
 _NEON2SSE_INLINE int32x4_t vornq_s32(int32x4_t a, int32x4_t b) // VORN q0,q0,q0
 {
     __m128i b1;
@@ -15498,7 +15500,7 @@ _NEON2SSE_INLINE int32x4_t vornq_s32(int32x4_t a, int32x4_t b) // VORN q0,q0,q0
     return _mm_or_si128 (a, b1);
 }
 
-int64x2_t vornq_s64(int64x2_t a, int64x2_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vornq_s64(int64x2_t a, int64x2_t b); // VORN q0,q0,q0
 _NEON2SSE_INLINE int64x2_t vornq_s64(int64x2_t a, int64x2_t b)
 {
     __m128i c1, b1;
@@ -15507,7 +15509,7 @@ _NEON2SSE_INLINE int64x2_t vornq_s64(int64x2_t a, int64x2_t b)
     return _mm_or_si128 (a, b1);
 }
 
-uint8x16_t vornq_u8(uint8x16_t a, uint8x16_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vornq_u8(uint8x16_t a, uint8x16_t b); // VORN q0,q0,q0
 _NEON2SSE_INLINE uint8x16_t vornq_u8(uint8x16_t a, uint8x16_t b) // VORN q0,q0,q0
 {
     __m128i b1;
@@ -15515,7 +15517,7 @@ _NEON2SSE_INLINE uint8x16_t vornq_u8(uint8x16_t a, uint8x16_t b) // VORN q0,q0,q
     return _mm_or_si128 (a, b1);
 }
 
-uint16x8_t vornq_u16(uint16x8_t a, uint16x8_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vornq_u16(uint16x8_t a, uint16x8_t b); // VORN q0,q0,q0
 _NEON2SSE_INLINE uint16x8_t vornq_u16(uint16x8_t a, uint16x8_t b) // VORN q0,q0,q0
 {
     __m128i b1;
@@ -15523,14 +15525,14 @@ _NEON2SSE_INLINE uint16x8_t vornq_u16(uint16x8_t a, uint16x8_t b) // VORN q0,q0,
     return _mm_or_si128 (a, b1);
 }
 
-uint32x4_t vornq_u32(uint32x4_t a, uint32x4_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vornq_u32(uint32x4_t a, uint32x4_t b); // VORN q0,q0,q0
 _NEON2SSE_INLINE uint32x4_t vornq_u32(uint32x4_t a, uint32x4_t b) // VORN q0,q0,q0
 {
     __m128i b1;
     b1 = vmvnq_u32( b); //bitwise not for b
     return _mm_or_si128 (a, b1);
 }
-uint64x2_t vornq_u64(uint64x2_t a, uint64x2_t b); // VORN q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vornq_u64(uint64x2_t a, uint64x2_t b); // VORN q0,q0,q0
 #define vornq_u64 vornq_s64
 
 //********************* Bitwise Select *****************************
@@ -15547,7 +15549,7 @@ uint64x2_t vornq_u64(uint64x2_t a, uint64x2_t b); // VORN q0,q0,q0
 //if the corresponding bit of the second operand is 1, otherwise leaves the destination bit unchanged.
 
 //VBSL only is implemented for SIMD
-int8x8_t vbsl_s8(uint8x8_t a, int8x8_t b, int8x8_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE int8x8_t vbsl_s8(uint8x8_t a, int8x8_t b, int8x8_t c); // VBSL d0,d0,d0
 _NEON2SSE_INLINE int8x8_t vbsl_s8(uint8x8_t a, int8x8_t b, int8x8_t c)
 {
     int8x8_t res64;
@@ -15556,13 +15558,13 @@ _NEON2SSE_INLINE int8x8_t vbsl_s8(uint8x8_t a, int8x8_t b, int8x8_t c)
     return64(res);
 }
 
-int16x4_t vbsl_s16(uint16x4_t a, int16x4_t b, int16x4_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE int16x4_t vbsl_s16(uint16x4_t a, int16x4_t b, int16x4_t c); // VBSL d0,d0,d0
 #define vbsl_s16 vbsl_s8
 
-int32x2_t vbsl_s32(uint32x2_t a, int32x2_t b, int32x2_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE int32x2_t vbsl_s32(uint32x2_t a, int32x2_t b, int32x2_t c); // VBSL d0,d0,d0
 #define vbsl_s32 vbsl_s8
 
-int64x1_t vbsl_s64(uint64x1_t a, int64x1_t b, int64x1_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE int64x1_t vbsl_s64(uint64x1_t a, int64x1_t b, int64x1_t c); // VBSL d0,d0,d0
 _NEON2SSE_INLINE int64x1_t vbsl_s64(uint64x1_t a, int64x1_t b, int64x1_t c)
 {
     int64x1_t res;
@@ -15570,19 +15572,19 @@ _NEON2SSE_INLINE int64x1_t vbsl_s64(uint64x1_t a, int64x1_t b, int64x1_t c)
     return res;
 }
 
-uint8x8_t vbsl_u8(uint8x8_t a,  uint8x8_t b, uint8x8_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE uint8x8_t vbsl_u8(uint8x8_t a,  uint8x8_t b, uint8x8_t c); // VBSL d0,d0,d0
 #define vbsl_u8 vbsl_s8
 
-uint16x4_t vbsl_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE uint16x4_t vbsl_u16(uint16x4_t a,  uint16x4_t b, uint16x4_t c); // VBSL d0,d0,d0
 #define vbsl_u16 vbsl_s8
 
-uint32x2_t vbsl_u32(uint32x2_t a,  uint32x2_t b, uint32x2_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE uint32x2_t vbsl_u32(uint32x2_t a,  uint32x2_t b, uint32x2_t c); // VBSL d0,d0,d0
 #define vbsl_u32 vbsl_s8
 
-uint64x1_t vbsl_u64(uint64x1_t a, uint64x1_t b, uint64x1_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE uint64x1_t vbsl_u64(uint64x1_t a, uint64x1_t b, uint64x1_t c); // VBSL d0,d0,d0
 #define vbsl_u64 vbsl_s64
 
-float32x2_t vbsl_f32(uint32x2_t a, float32x2_t b, float32x2_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE float32x2_t vbsl_f32(uint32x2_t a, float32x2_t b, float32x2_t c); // VBSL d0,d0,d0
 _NEON2SSE_INLINE float32x2_t vbsl_f32(uint32x2_t a, float32x2_t b, float32x2_t c)
 {
     __m128 sel1, sel2;
@@ -15594,13 +15596,13 @@ _NEON2SSE_INLINE float32x2_t vbsl_f32(uint32x2_t a, float32x2_t b, float32x2_t c
     return res64;
 }
 
-poly8x8_t vbsl_p8(uint8x8_t a, poly8x8_t b, poly8x8_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE poly8x8_t vbsl_p8(uint8x8_t a, poly8x8_t b, poly8x8_t c); // VBSL d0,d0,d0
 #define  vbsl_p8 vbsl_s8
 
-poly16x4_t vbsl_p16(uint16x4_t a, poly16x4_t b, poly16x4_t c); // VBSL d0,d0,d0
+_NEON2SSESTORAGE poly16x4_t vbsl_p16(uint16x4_t a, poly16x4_t b, poly16x4_t c); // VBSL d0,d0,d0
 #define  vbsl_p16 vbsl_s8
 
-int8x16_t vbslq_s8(uint8x16_t a, int8x16_t b, int8x16_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE int8x16_t vbslq_s8(uint8x16_t a, int8x16_t b, int8x16_t c); // VBSL q0,q0,q0
 _NEON2SSE_INLINE int8x16_t vbslq_s8(uint8x16_t a, int8x16_t b, int8x16_t c) // VBSL q0,q0,q0
 {
     __m128i sel1, sel2;
@@ -15609,28 +15611,28 @@ _NEON2SSE_INLINE int8x16_t vbslq_s8(uint8x16_t a, int8x16_t b, int8x16_t c) // V
     return _mm_or_si128 (sel1, sel2);
 }
 
-int16x8_t vbslq_s16(uint16x8_t a, int16x8_t b, int16x8_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE int16x8_t vbslq_s16(uint16x8_t a, int16x8_t b, int16x8_t c); // VBSL q0,q0,q0
 #define vbslq_s16 vbslq_s8
 
-int32x4_t vbslq_s32(uint32x4_t a, int32x4_t b, int32x4_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE int32x4_t vbslq_s32(uint32x4_t a, int32x4_t b, int32x4_t c); // VBSL q0,q0,q0
 #define vbslq_s32 vbslq_s8
 
-int64x2_t vbslq_s64(uint64x2_t a, int64x2_t b, int64x2_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE int64x2_t vbslq_s64(uint64x2_t a, int64x2_t b, int64x2_t c); // VBSL q0,q0,q0
 #define vbslq_s64 vbslq_s8
 
-uint8x16_t vbslq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE uint8x16_t vbslq_u8(uint8x16_t a, uint8x16_t b, uint8x16_t c); // VBSL q0,q0,q0
 #define vbslq_u8 vbslq_s8
 
-uint16x8_t vbslq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE uint16x8_t vbslq_u16(uint16x8_t a, uint16x8_t b, uint16x8_t c); // VBSL q0,q0,q0
 #define vbslq_u16 vbslq_s8
 
-uint32x4_t vbslq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE uint32x4_t vbslq_u32(uint32x4_t a, uint32x4_t b, uint32x4_t c); // VBSL q0,q0,q0
 #define vbslq_u32 vbslq_s8
 
-uint64x2_t vbslq_u64(uint64x2_t a, uint64x2_t b, uint64x2_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE uint64x2_t vbslq_u64(uint64x2_t a, uint64x2_t b, uint64x2_t c); // VBSL q0,q0,q0
 #define vbslq_u64 vbslq_s8
 
-float32x4_t vbslq_f32(uint32x4_t a, float32x4_t b, float32x4_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE float32x4_t vbslq_f32(uint32x4_t a, float32x4_t b, float32x4_t c); // VBSL q0,q0,q0
 _NEON2SSE_INLINE float32x4_t vbslq_f32(uint32x4_t a, float32x4_t b, float32x4_t c) // VBSL q0,q0,q0
 {
     __m128 sel1, sel2;
@@ -15639,10 +15641,10 @@ _NEON2SSE_INLINE float32x4_t vbslq_f32(uint32x4_t a, float32x4_t b, float32x4_t
     return _mm_or_ps (sel1, sel2);
 }
 
-poly8x16_t vbslq_p8(uint8x16_t a, poly8x16_t b, poly8x16_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE poly8x16_t vbslq_p8(uint8x16_t a, poly8x16_t b, poly8x16_t c); // VBSL q0,q0,q0
 #define vbslq_p8 vbslq_u8
 
-poly16x8_t vbslq_p16(uint16x8_t a, poly16x8_t b, poly16x8_t c); // VBSL q0,q0,q0
+_NEON2SSESTORAGE poly16x8_t vbslq_p16(uint16x8_t a, poly16x8_t b, poly16x8_t c); // VBSL q0,q0,q0
 #define vbslq_p16 vbslq_s8
 
 //************************************************************************************
@@ -15652,7 +15654,7 @@ poly16x8_t vbslq_p16(uint16x8_t a, poly16x8_t b, poly16x8_t c); // VBSL q0,q0,q0
 //************************************************************************************
 //VTRN (Vector Transpose) treats the elements of its operand vectors as elements of 2 x 2 matrices, and transposes the matrices.
 // making the result look as (a0, b0, a2, b2, a4, b4,....) (a1, b1, a3, b3, a5, b5,.....)
-int8x8x2_t vtrn_s8(int8x8_t a, int8x8_t b); // VTRN.8 d0,d0
+_NEON2SSESTORAGE int8x8x2_t vtrn_s8(int8x8_t a, int8x8_t b); // VTRN.8 d0,d0
 _NEON2SSE_INLINE int8x8x2_t vtrn_s8(int8x8_t a, int8x8_t b) // VTRN.8 d0,d0
 {
     int8x8x2_t val;
@@ -15663,19 +15665,19 @@ _NEON2SSE_INLINE int8x8x2_t vtrn_s8(int8x8_t a, int8x8_t b) // VTRN.8 d0,d0
     return val;
 }
 
-int16x4x2_t vtrn_s16(int16x4_t a, int16x4_t b); // VTRN.16 d0,d0
+_NEON2SSESTORAGE int16x4x2_t vtrn_s16(int16x4_t a, int16x4_t b); // VTRN.16 d0,d0
 _NEON2SSE_INLINE int16x4x2_t vtrn_s16(int16x4_t a, int16x4_t b) // VTRN.16 d0,d0
 {
     int16x4x2_t val;
     __m128i tmp, val0;
-    _NEON2SSE_ALIGN_16 int8_t maskdlv16[16] = {0,1, 2,3, 8,9, 10,11, 4,5, 6,7, 12,13, 14, 15};
+    _NEON2SSE_ALIGN_16 static const int8_t maskdlv16[16] = {0,1, 2,3, 8,9, 10,11, 4,5, 6,7, 12,13, 14, 15};
     tmp = _mm_unpacklo_epi16(_pM128i(a), _pM128i(b)); //a0,b0,a1,b1,a2,b2,a3,b3
     val0 =  _mm_shuffle_epi8 (tmp, *(__m128i*)maskdlv16); //a0, b0, a2, b2, a1,b1, a3, b3
     vst1q_s16(val.val, val0); // _mm_shuffle_epi32 (val.val[0], _SWAP_HI_LOW32); //(a1,b1, a3,b3),(a0, b0, a2, b2),
     return val;
 }
 
-int32x2x2_t vtrn_s32(int32x2_t a, int32x2_t b); // VTRN.32 d0,d0
+_NEON2SSESTORAGE int32x2x2_t vtrn_s32(int32x2_t a, int32x2_t b); // VTRN.32 d0,d0
 _NEON2SSE_INLINE int32x2x2_t vtrn_s32(int32x2_t a, int32x2_t b)
 {
     int32x2x2_t val;
@@ -15685,16 +15687,16 @@ _NEON2SSE_INLINE int32x2x2_t vtrn_s32(int32x2_t a, int32x2_t b)
     return val;
 }
 
-uint8x8x2_t vtrn_u8(uint8x8_t a, uint8x8_t b); // VTRN.8 d0,d0
+_NEON2SSESTORAGE uint8x8x2_t vtrn_u8(uint8x8_t a, uint8x8_t b); // VTRN.8 d0,d0
 #define vtrn_u8 vtrn_s8
 
-uint16x4x2_t vtrn_u16(uint16x4_t a, uint16x4_t b); // VTRN.16 d0,d0
+_NEON2SSESTORAGE uint16x4x2_t vtrn_u16(uint16x4_t a, uint16x4_t b); // VTRN.16 d0,d0
 #define vtrn_u16 vtrn_s16
 
-uint32x2x2_t vtrn_u32(uint32x2_t a, uint32x2_t b); // VTRN.32 d0,d0
+_NEON2SSESTORAGE uint32x2x2_t vtrn_u32(uint32x2_t a, uint32x2_t b); // VTRN.32 d0,d0
 #define vtrn_u32 vtrn_s32
 
-float32x2x2_t vtrn_f32(float32x2_t a, float32x2_t b); // VTRN.32 d0,d0
+_NEON2SSESTORAGE float32x2x2_t vtrn_f32(float32x2_t a, float32x2_t b); // VTRN.32 d0,d0
 _NEON2SSE_INLINE float32x2x2_t vtrn_f32(float32x2_t a, float32x2_t b)
 {
     float32x2x2_t val;
@@ -15705,10 +15707,10 @@ _NEON2SSE_INLINE float32x2x2_t vtrn_f32(float32x2_t a, float32x2_t b)
     return val; //a0,b0,a1,b1
 }
 
-poly8x8x2_t vtrn_p8(poly8x8_t a, poly8x8_t b); // VTRN.8 d0,d0
+_NEON2SSESTORAGE poly8x8x2_t vtrn_p8(poly8x8_t a, poly8x8_t b); // VTRN.8 d0,d0
 #define  vtrn_p8 vtrn_u8
 
-poly16x4x2_t vtrn_p16(poly16x4_t a, poly16x4_t b); // VTRN.16 d0,d0
+_NEON2SSESTORAGE poly16x4x2_t vtrn_p16(poly16x4_t a, poly16x4_t b); // VTRN.16 d0,d0
 #define  vtrn_p16 vtrn_s16
 
 //int8x16x2_t vtrnq_s8(int8x16_t a, int8x16_t b); // VTRN.8 q0,q0
@@ -15724,7 +15726,7 @@ _NEON2SSE_INLINE int8x16x2_t vtrnq_s8(int8x16_t a, int8x16_t b) // VTRN.8 q0,q0
     return r8x16;
 }
 
-int16x8x2_t vtrnq_s16(int16x8_t a, int16x8_t b); // VTRN.16 q0,q0
+_NEON2SSESTORAGE int16x8x2_t vtrnq_s16(int16x8_t a, int16x8_t b); // VTRN.16 q0,q0
 _NEON2SSE_INLINE int16x8x2_t vtrnq_s16(int16x8_t a, int16x8_t b) // VTRN.16 q0,q0
 {
     int16x8x2_t v16x8;
@@ -15736,7 +15738,7 @@ _NEON2SSE_INLINE int16x8x2_t vtrnq_s16(int16x8_t a, int16x8_t b) // VTRN.16 q0,q
     return v16x8;
 }
 
-int32x4x2_t vtrnq_s32(int32x4_t a, int32x4_t b); // VTRN.32 q0,q0
+_NEON2SSESTORAGE int32x4x2_t vtrnq_s32(int32x4_t a, int32x4_t b); // VTRN.32 q0,q0
 _NEON2SSE_INLINE int32x4x2_t vtrnq_s32(int32x4_t a, int32x4_t b) // VTRN.32 q0,q0
 {
     //may be not optimal solution compared with serial
@@ -15750,16 +15752,16 @@ _NEON2SSE_INLINE int32x4x2_t vtrnq_s32(int32x4_t a, int32x4_t b) // VTRN.32 q0,q
     return v32x4;
 }
 
-uint8x16x2_t vtrnq_u8(uint8x16_t a, uint8x16_t b); // VTRN.8 q0,q0
+_NEON2SSESTORAGE uint8x16x2_t vtrnq_u8(uint8x16_t a, uint8x16_t b); // VTRN.8 q0,q0
 #define vtrnq_u8 vtrnq_s8
 
-uint16x8x2_t vtrnq_u16(uint16x8_t a, uint16x8_t b); // VTRN.16 q0,q0
+_NEON2SSESTORAGE uint16x8x2_t vtrnq_u16(uint16x8_t a, uint16x8_t b); // VTRN.16 q0,q0
 #define vtrnq_u16 vtrnq_s16
 
-uint32x4x2_t vtrnq_u32(uint32x4_t a, uint32x4_t b); // VTRN.32 q0,q0
+_NEON2SSESTORAGE uint32x4x2_t vtrnq_u32(uint32x4_t a, uint32x4_t b); // VTRN.32 q0,q0
 #define vtrnq_u32 vtrnq_s32
 
-float32x4x2_t vtrnq_f32(float32x4_t a, float32x4_t b); // VTRN.32 q0,q0
+_NEON2SSESTORAGE float32x4x2_t vtrnq_f32(float32x4_t a, float32x4_t b); // VTRN.32 q0,q0
 _NEON2SSE_INLINE float32x4x2_t vtrnq_f32(float32x4_t a, float32x4_t b) // VTRN.32 q0,q0
 {
     //may be not optimal solution compared with serial
@@ -15773,16 +15775,16 @@ _NEON2SSE_INLINE float32x4x2_t vtrnq_f32(float32x4_t a, float32x4_t b) // VTRN.3
     return f32x4;
 }
 
-poly8x16x2_t vtrnq_p8(poly8x16_t a, poly8x16_t b); // VTRN.8 q0,q0
+_NEON2SSESTORAGE poly8x16x2_t vtrnq_p8(poly8x16_t a, poly8x16_t b); // VTRN.8 q0,q0
 #define vtrnq_p8 vtrnq_s8
 
-poly16x8x2_t vtrnq_p16(poly16x8_t a, poly16x8_t b); // VTRN.16 q0,q0
+_NEON2SSESTORAGE poly16x8x2_t vtrnq_p16(poly16x8_t a, poly16x8_t b); // VTRN.16 q0,q0
 #define vtrnq_p16 vtrnq_s16
 
 //***************** Interleave elements ***************************
 //*****************************************************************
 //output has (a0,b0,a1,b1, a2,b2,.....)
-int8x8x2_t vzip_s8(int8x8_t a, int8x8_t b); // VZIP.8 d0,d0
+_NEON2SSESTORAGE int8x8x2_t vzip_s8(int8x8_t a, int8x8_t b); // VZIP.8 d0,d0
 _NEON2SSE_INLINE int8x8x2_t vzip_s8(int8x8_t a, int8x8_t b) // VZIP.8 d0,d0
 {
     int8x8x2_t val;
@@ -15792,7 +15794,7 @@ _NEON2SSE_INLINE int8x8x2_t vzip_s8(int8x8_t a, int8x8_t b) // VZIP.8 d0,d0
     return val;
 }
 
-int16x4x2_t vzip_s16(int16x4_t a, int16x4_t b); // VZIP.16 d0,d0
+_NEON2SSESTORAGE int16x4x2_t vzip_s16(int16x4_t a, int16x4_t b); // VZIP.16 d0,d0
 _NEON2SSE_INLINE int16x4x2_t vzip_s16(int16x4_t a, int16x4_t b) // VZIP.16 d0,d0
 {
     int16x4x2_t val;
@@ -15802,28 +15804,28 @@ _NEON2SSE_INLINE int16x4x2_t vzip_s16(int16x4_t a, int16x4_t b) // VZIP.16 d0,d0
     return val;
 }
 
-int32x2x2_t vzip_s32(int32x2_t a, int32x2_t b); // VZIP.32 d0,d0
+_NEON2SSESTORAGE int32x2x2_t vzip_s32(int32x2_t a, int32x2_t b); // VZIP.32 d0,d0
 #define vzip_s32 vtrn_s32
 
-uint8x8x2_t vzip_u8(uint8x8_t a, uint8x8_t b); // VZIP.8 d0,d0
+_NEON2SSESTORAGE uint8x8x2_t vzip_u8(uint8x8_t a, uint8x8_t b); // VZIP.8 d0,d0
 #define vzip_u8 vzip_s8
 
-uint16x4x2_t vzip_u16(uint16x4_t a, uint16x4_t b); // VZIP.16 d0,d0
+_NEON2SSESTORAGE uint16x4x2_t vzip_u16(uint16x4_t a, uint16x4_t b); // VZIP.16 d0,d0
 #define vzip_u16 vzip_s16
 
-uint32x2x2_t vzip_u32(uint32x2_t a, uint32x2_t b); // VZIP.32 d0,d0
+_NEON2SSESTORAGE uint32x2x2_t vzip_u32(uint32x2_t a, uint32x2_t b); // VZIP.32 d0,d0
 #define vzip_u32 vzip_s32
 
-float32x2x2_t vzip_f32(float32x2_t a, float32x2_t b); // VZIP.32 d0,d0
+_NEON2SSESTORAGE float32x2x2_t vzip_f32(float32x2_t a, float32x2_t b); // VZIP.32 d0,d0
 #define vzip_f32 vtrn_f32
 
-poly8x8x2_t vzip_p8(poly8x8_t a, poly8x8_t b); // VZIP.8 d0,d0
+_NEON2SSESTORAGE poly8x8x2_t vzip_p8(poly8x8_t a, poly8x8_t b); // VZIP.8 d0,d0
 #define vzip_p8 vzip_u8
 
-poly16x4x2_t vzip_p16(poly16x4_t a, poly16x4_t b); // VZIP.16 d0,d0
+_NEON2SSESTORAGE poly16x4x2_t vzip_p16(poly16x4_t a, poly16x4_t b); // VZIP.16 d0,d0
 #define vzip_p16 vzip_u16
 
-int8x16x2_t vzipq_s8(int8x16_t a, int8x16_t b); // VZIP.8 q0,q0
+_NEON2SSESTORAGE int8x16x2_t vzipq_s8(int8x16_t a, int8x16_t b); // VZIP.8 q0,q0
 _NEON2SSE_INLINE int8x16x2_t vzipq_s8(int8x16_t a, int8x16_t b) // VZIP.8 q0,q0
 {
     int8x16x2_t r8x16;
@@ -15832,7 +15834,7 @@ _NEON2SSE_INLINE int8x16x2_t vzipq_s8(int8x16_t a, int8x16_t b) // VZIP.8 q0,q0
     return r8x16;
 }
 
-int16x8x2_t vzipq_s16(int16x8_t a, int16x8_t b); // VZIP.16 q0,q0
+_NEON2SSESTORAGE int16x8x2_t vzipq_s16(int16x8_t a, int16x8_t b); // VZIP.16 q0,q0
 _NEON2SSE_INLINE int16x8x2_t vzipq_s16(int16x8_t a, int16x8_t b) // VZIP.16 q0,q0
 {
     int16x8x2_t r16x8;
@@ -15841,7 +15843,7 @@ _NEON2SSE_INLINE int16x8x2_t vzipq_s16(int16x8_t a, int16x8_t b) // VZIP.16 q0,q
     return r16x8;
 }
 
-int32x4x2_t vzipq_s32(int32x4_t a, int32x4_t b); // VZIP.32 q0,q0
+_NEON2SSESTORAGE int32x4x2_t vzipq_s32(int32x4_t a, int32x4_t b); // VZIP.32 q0,q0
 _NEON2SSE_INLINE int32x4x2_t vzipq_s32(int32x4_t a, int32x4_t b) // VZIP.32 q0,q0
 {
     int32x4x2_t r32x4;
@@ -15850,16 +15852,16 @@ _NEON2SSE_INLINE int32x4x2_t vzipq_s32(int32x4_t a, int32x4_t b) // VZIP.32 q0,q
     return r32x4;
 }
 
-uint8x16x2_t vzipq_u8(uint8x16_t a, uint8x16_t b); // VZIP.8 q0,q0
+_NEON2SSESTORAGE uint8x16x2_t vzipq_u8(uint8x16_t a, uint8x16_t b); // VZIP.8 q0,q0
 #define vzipq_u8 vzipq_s8
 
-uint16x8x2_t vzipq_u16(uint16x8_t a, uint16x8_t b); // VZIP.16 q0,q0
+_NEON2SSESTORAGE uint16x8x2_t vzipq_u16(uint16x8_t a, uint16x8_t b); // VZIP.16 q0,q0
 #define vzipq_u16 vzipq_s16
 
-uint32x4x2_t vzipq_u32(uint32x4_t a, uint32x4_t b); // VZIP.32 q0,q0
+_NEON2SSESTORAGE uint32x4x2_t vzipq_u32(uint32x4_t a, uint32x4_t b); // VZIP.32 q0,q0
 #define vzipq_u32 vzipq_s32
 
-float32x4x2_t vzipq_f32(float32x4_t a, float32x4_t b); // VZIP.32 q0,q0
+_NEON2SSESTORAGE float32x4x2_t vzipq_f32(float32x4_t a, float32x4_t b); // VZIP.32 q0,q0
 _NEON2SSE_INLINE float32x4x2_t vzipq_f32(float32x4_t a, float32x4_t b) // VZIP.32 q0,q0
 {
     float32x4x2_t f32x4;
@@ -15868,41 +15870,41 @@ _NEON2SSE_INLINE float32x4x2_t vzipq_f32(float32x4_t a, float32x4_t b) // VZIP.3
     return f32x4;
 }
 
-poly8x16x2_t vzipq_p8(poly8x16_t a, poly8x16_t b); // VZIP.8 q0,q0
+_NEON2SSESTORAGE poly8x16x2_t vzipq_p8(poly8x16_t a, poly8x16_t b); // VZIP.8 q0,q0
 #define vzipq_p8 vzipq_u8
 
-poly16x8x2_t vzipq_p16(poly16x8_t a, poly16x8_t b); // VZIP.16 q0,q0
+_NEON2SSESTORAGE poly16x8x2_t vzipq_p16(poly16x8_t a, poly16x8_t b); // VZIP.16 q0,q0
 #define vzipq_p16 vzipq_u16
 
 //*********************** De-Interleave elements *************************
 //*************************************************************************
 //As the result of these functions first val  contains (a0,a2,a4,....,b0,b2, b4,...) and the second val (a1,a3,a5,....b1,b3,b5...)
 //no such functions in IA32 SIMD, shuffle is required
-int8x8x2_t vuzp_s8(int8x8_t a, int8x8_t b); // VUZP.8 d0,d0
+_NEON2SSESTORAGE int8x8x2_t vuzp_s8(int8x8_t a, int8x8_t b); // VUZP.8 d0,d0
 _NEON2SSE_INLINE int8x8x2_t vuzp_s8(int8x8_t a, int8x8_t b) // VUZP.8 d0,d0
 {
     int8x8x2_t val;
     __m128i tmp, val0;
-    _NEON2SSE_ALIGN_16 int8_t maskdlv8[16] = { 0, 4, 8, 12, 1, 5, 9, 13,  2, 6, 10, 14, 3, 7, 11,15};
+    _NEON2SSE_ALIGN_16 static const int8_t maskdlv8[16] = { 0, 4, 8, 12, 1, 5, 9, 13,  2, 6, 10, 14, 3, 7, 11,15};
     tmp = _mm_unpacklo_epi8(_pM128i(a), _pM128i(b)); //a0,b0,a1,b1,a2,b2,a3,b3,...,a7,b7
     val0 = _mm_shuffle_epi8 (tmp, *(__m128i*)maskdlv8); //(a0, a2, a4, a6, b0, b2, b4, b6),  (a1, a3, a5, a7, b1,b3, b5, b7)
     vst1q_s8(val.val, val0); // _mm_shuffle_epi32(val.val[0], _SWAP_HI_LOW32);
     return val;
 }
 
-int16x4x2_t vuzp_s16(int16x4_t a, int16x4_t b); // VUZP.16 d0,d0
+_NEON2SSESTORAGE int16x4x2_t vuzp_s16(int16x4_t a, int16x4_t b); // VUZP.16 d0,d0
 _NEON2SSE_INLINE int16x4x2_t vuzp_s16(int16x4_t a, int16x4_t b) // VUZP.16 d0,d0
 {
     int16x4x2_t val;
     __m128i tmp, val0;
-    _NEON2SSE_ALIGN_16 int8_t maskdlv16[16] = {0,1,  8,9,  2,3, 10,11,  4,5, 12,13, 6,7, 14,15};
+    _NEON2SSE_ALIGN_16 static const int8_t maskdlv16[16] = {0,1,  8,9,  2,3, 10,11,  4,5, 12,13, 6,7, 14,15};
     tmp = _mm_unpacklo_epi16(_pM128i(a), _pM128i(b)); //a0,b0,a1,b1,a2,b2,a3,b3
     val0 = _mm_shuffle_epi8 (tmp, *(__m128i*)maskdlv16); //a0,a2, b0, b2, a1,a3, b1,b3
     vst1q_s16(val.val, val0); // _mm_shuffle_epi32(val.val[0], _SWAP_HI_LOW32);
     return val;
 }
 
-int32x2x2_t vuzp_s32(int32x2_t a, int32x2_t b); // VUZP.32 d0,d0
+_NEON2SSESTORAGE int32x2x2_t vuzp_s32(int32x2_t a, int32x2_t b); // VUZP.32 d0,d0
 _NEON2SSE_INLINE int32x2x2_t vuzp_s32(int32x2_t a, int32x2_t b) // VUZP.32 d0,d0
 {
     int32x2x2_t val;
@@ -15912,25 +15914,25 @@ _NEON2SSE_INLINE int32x2x2_t vuzp_s32(int32x2_t a, int32x2_t b) // VUZP.32 d0,d0
     return val;
 }
 
-uint8x8x2_t vuzp_u8(uint8x8_t a, uint8x8_t b); // VUZP.8 d0,d0
+_NEON2SSESTORAGE uint8x8x2_t vuzp_u8(uint8x8_t a, uint8x8_t b); // VUZP.8 d0,d0
 #define vuzp_u8 vuzp_s8
 
-uint16x4x2_t vuzp_u16(uint16x4_t a, uint16x4_t b); // VUZP.16 d0,d0
+_NEON2SSESTORAGE uint16x4x2_t vuzp_u16(uint16x4_t a, uint16x4_t b); // VUZP.16 d0,d0
 #define vuzp_u16 vuzp_s16
 
-uint32x2x2_t vuzp_u32(uint32x2_t a, uint32x2_t b); // VUZP.32 d0,d0
+_NEON2SSESTORAGE uint32x2x2_t vuzp_u32(uint32x2_t a, uint32x2_t b); // VUZP.32 d0,d0
 #define vuzp_u32 vuzp_s32
 
-float32x2x2_t vuzp_f32(float32x2_t a, float32x2_t b); // VUZP.32 d0,d0
+_NEON2SSESTORAGE float32x2x2_t vuzp_f32(float32x2_t a, float32x2_t b); // VUZP.32 d0,d0
 #define vuzp_f32 vzip_f32
 
-poly8x8x2_t vuzp_p8(poly8x8_t a, poly8x8_t b); // VUZP.8 d0,d0
+_NEON2SSESTORAGE poly8x8x2_t vuzp_p8(poly8x8_t a, poly8x8_t b); // VUZP.8 d0,d0
 #define vuzp_p8 vuzp_u8
 
-poly16x4x2_t vuzp_p16(poly16x4_t a, poly16x4_t b); // VUZP.16 d0,d0
+_NEON2SSESTORAGE poly16x4x2_t vuzp_p16(poly16x4_t a, poly16x4_t b); // VUZP.16 d0,d0
 #define vuzp_p16 vuzp_u16
 
-int8x16x2_t vuzpq_s8(int8x16_t a, int8x16_t b); // VUZP.8 q0,q0
+_NEON2SSESTORAGE int8x16x2_t vuzpq_s8(int8x16_t a, int8x16_t b); // VUZP.8 q0,q0
 _NEON2SSE_INLINE int8x16x2_t vuzpq_s8(int8x16_t a, int8x16_t b) // VUZP.8 q0,q0
 {
     int8x16x2_t v8x16;
@@ -15943,7 +15945,7 @@ _NEON2SSE_INLINE int8x16x2_t vuzpq_s8(int8x16_t a, int8x16_t b) // VUZP.8 q0,q0
     return v8x16;
 }
 
-int16x8x2_t vuzpq_s16(int16x8_t a, int16x8_t b); // VUZP.16 q0,q0
+_NEON2SSESTORAGE int16x8x2_t vuzpq_s16(int16x8_t a, int16x8_t b); // VUZP.16 q0,q0
 _NEON2SSE_INLINE int16x8x2_t vuzpq_s16(int16x8_t a, int16x8_t b) // VUZP.16 q0,q0
 {
     int16x8x2_t v16x8;
@@ -15955,7 +15957,7 @@ _NEON2SSE_INLINE int16x8x2_t vuzpq_s16(int16x8_t a, int16x8_t b) // VUZP.16 q0,q
     return v16x8;
 }
 
-int32x4x2_t vuzpq_s32(int32x4_t a, int32x4_t b); // VUZP.32 q0,q0
+_NEON2SSESTORAGE int32x4x2_t vuzpq_s32(int32x4_t a, int32x4_t b); // VUZP.32 q0,q0
 _NEON2SSE_INLINE int32x4x2_t vuzpq_s32(int32x4_t a, int32x4_t b) // VUZP.32 q0,q0
 {
     //may be not optimal solution compared with serial
@@ -15969,16 +15971,16 @@ _NEON2SSE_INLINE int32x4x2_t vuzpq_s32(int32x4_t a, int32x4_t b) // VUZP.32 q0,q
     return v32x4;
 }
 
-uint8x16x2_t vuzpq_u8(uint8x16_t a, uint8x16_t b); // VUZP.8 q0,q0
+_NEON2SSESTORAGE uint8x16x2_t vuzpq_u8(uint8x16_t a, uint8x16_t b); // VUZP.8 q0,q0
 #define vuzpq_u8 vuzpq_s8
 
-uint16x8x2_t vuzpq_u16(uint16x8_t a, uint16x8_t b); // VUZP.16 q0,q0
+_NEON2SSESTORAGE uint16x8x2_t vuzpq_u16(uint16x8_t a, uint16x8_t b); // VUZP.16 q0,q0
 #define vuzpq_u16 vuzpq_s16
 
-uint32x4x2_t vuzpq_u32(uint32x4_t a, uint32x4_t b); // VUZP.32 q0,q0
+_NEON2SSESTORAGE uint32x4x2_t vuzpq_u32(uint32x4_t a, uint32x4_t b); // VUZP.32 q0,q0
 #define vuzpq_u32 vuzpq_s32
 
-float32x4x2_t vuzpq_f32(float32x4_t a, float32x4_t b); // VUZP.32 q0,q0
+_NEON2SSESTORAGE float32x4x2_t vuzpq_f32(float32x4_t a, float32x4_t b); // VUZP.32 q0,q0
 _NEON2SSE_INLINE float32x4x2_t vuzpq_f32(float32x4_t a, float32x4_t b) // VUZP.32 q0,q0
 {
     float32x4x2_t v32x4;
@@ -15987,691 +15989,691 @@ _NEON2SSE_INLINE float32x4x2_t vuzpq_f32(float32x4_t a, float32x4_t b) // VUZP.3
     return v32x4;
 }
 
-poly8x16x2_t vuzpq_p8(poly8x16_t a, poly8x16_t b); // VUZP.8 q0,q0
+_NEON2SSESTORAGE poly8x16x2_t vuzpq_p8(poly8x16_t a, poly8x16_t b); // VUZP.8 q0,q0
 #define vuzpq_p8 vuzpq_u8
 
-poly16x8x2_t vuzpq_p16(poly16x8_t a, poly16x8_t b); // VUZP.16 q0,q0
+_NEON2SSESTORAGE poly16x8x2_t vuzpq_p16(poly16x8_t a, poly16x8_t b); // VUZP.16 q0,q0
 #define vuzpq_p16 vuzpq_u16
 
 //##############################################################################################
 //*********************** Reinterpret cast intrinsics.******************************************
 //##############################################################################################
 // Not a part of oficial NEON instruction set but available in gcc compiler *********************
-poly8x8_t vreinterpret_p8_u32 (uint32x2_t t);
+_NEON2SSESTORAGE poly8x8_t vreinterpret_p8_u32 (uint32x2_t t);
 #define vreinterpret_p8_u32
 
-poly8x8_t vreinterpret_p8_u16 (uint16x4_t t);
+_NEON2SSESTORAGE poly8x8_t vreinterpret_p8_u16 (uint16x4_t t);
 #define vreinterpret_p8_u16
 
-poly8x8_t vreinterpret_p8_u8 (uint8x8_t t);
+_NEON2SSESTORAGE poly8x8_t vreinterpret_p8_u8 (uint8x8_t t);
 #define vreinterpret_p8_u8
 
-poly8x8_t vreinterpret_p8_s32 (int32x2_t t);
+_NEON2SSESTORAGE poly8x8_t vreinterpret_p8_s32 (int32x2_t t);
 #define vreinterpret_p8_s32
 
-poly8x8_t vreinterpret_p8_s16 (int16x4_t t);
+_NEON2SSESTORAGE poly8x8_t vreinterpret_p8_s16 (int16x4_t t);
 #define vreinterpret_p8_s16
 
-poly8x8_t vreinterpret_p8_s8 (int8x8_t t);
+_NEON2SSESTORAGE poly8x8_t vreinterpret_p8_s8 (int8x8_t t);
 #define vreinterpret_p8_s8
 
-poly8x8_t vreinterpret_p8_u64 (uint64x1_t t);
+_NEON2SSESTORAGE poly8x8_t vreinterpret_p8_u64 (uint64x1_t t);
 #define vreinterpret_p8_u64
 
-poly8x8_t vreinterpret_p8_s64 (int64x1_t t);
+_NEON2SSESTORAGE poly8x8_t vreinterpret_p8_s64 (int64x1_t t);
 #define vreinterpret_p8_s64
 
-poly8x8_t vreinterpret_p8_f32 (float32x2_t t);
+_NEON2SSESTORAGE poly8x8_t vreinterpret_p8_f32 (float32x2_t t);
 #define vreinterpret_p8_f32
 
-poly8x8_t vreinterpret_p8_p16 (poly16x4_t t);
+_NEON2SSESTORAGE poly8x8_t vreinterpret_p8_p16 (poly16x4_t t);
 #define vreinterpret_p8_p16
 
-poly8x16_t vreinterpretq_p8_u32 (uint32x4_t t);
+_NEON2SSESTORAGE poly8x16_t vreinterpretq_p8_u32 (uint32x4_t t);
 #define vreinterpretq_p8_u32
 
-poly8x16_t vreinterpretq_p8_u16 (uint16x8_t t);
+_NEON2SSESTORAGE poly8x16_t vreinterpretq_p8_u16 (uint16x8_t t);
 #define vreinterpretq_p8_u16
 
-poly8x16_t vreinterpretq_p8_u8 (uint8x16_t t);
+_NEON2SSESTORAGE poly8x16_t vreinterpretq_p8_u8 (uint8x16_t t);
 #define vreinterpretq_p8_u8
 
-poly8x16_t vreinterpretq_p8_s32 (int32x4_t t);
+_NEON2SSESTORAGE poly8x16_t vreinterpretq_p8_s32 (int32x4_t t);
 #define vreinterpretq_p8_s32
 
-poly8x16_t vreinterpretq_p8_s16 (int16x8_t t);
+_NEON2SSESTORAGE poly8x16_t vreinterpretq_p8_s16 (int16x8_t t);
 #define vreinterpretq_p8_s16
 
-poly8x16_t vreinterpretq_p8_s8 (int8x16_t t);
+_NEON2SSESTORAGE poly8x16_t vreinterpretq_p8_s8 (int8x16_t t);
 #define vreinterpretq_p8_s8
 
-poly8x16_t vreinterpretq_p8_u64 (uint64x2_t t);
+_NEON2SSESTORAGE poly8x16_t vreinterpretq_p8_u64 (uint64x2_t t);
 #define vreinterpretq_p8_u64
 
-poly8x16_t vreinterpretq_p8_s64 (int64x2_t t);
+_NEON2SSESTORAGE poly8x16_t vreinterpretq_p8_s64 (int64x2_t t);
 #define vreinterpretq_p8_s64
 
-poly8x16_t vreinterpretq_p8_f32 (float32x4_t t);
+_NEON2SSESTORAGE poly8x16_t vreinterpretq_p8_f32 (float32x4_t t);
 #define vreinterpretq_p8_f32(t) _M128i(t)
 
-poly8x16_t vreinterpretq_p8_p16 (poly16x8_t t);
+_NEON2SSESTORAGE poly8x16_t vreinterpretq_p8_p16 (poly16x8_t t);
 #define vreinterpretq_p8_p16
 
-poly16x4_t vreinterpret_p16_u32 (uint32x2_t t);
+_NEON2SSESTORAGE poly16x4_t vreinterpret_p16_u32 (uint32x2_t t);
 #define vreinterpret_p16_u32
 
-poly16x4_t vreinterpret_p16_u16 (uint16x4_t t);
+_NEON2SSESTORAGE poly16x4_t vreinterpret_p16_u16 (uint16x4_t t);
 #define vreinterpret_p16_u16
 
-poly16x4_t vreinterpret_p16_u8 (uint8x8_t t);
+_NEON2SSESTORAGE poly16x4_t vreinterpret_p16_u8 (uint8x8_t t);
 #define vreinterpret_p16_u8
 
-poly16x4_t vreinterpret_p16_s32 (int32x2_t t);
+_NEON2SSESTORAGE poly16x4_t vreinterpret_p16_s32 (int32x2_t t);
 #define vreinterpret_p16_s32
 
-poly16x4_t vreinterpret_p16_s16 (int16x4_t t);
+_NEON2SSESTORAGE poly16x4_t vreinterpret_p16_s16 (int16x4_t t);
 #define vreinterpret_p16_s16
 
-poly16x4_t vreinterpret_p16_s8 (int8x8_t t);
+_NEON2SSESTORAGE poly16x4_t vreinterpret_p16_s8 (int8x8_t t);
 #define vreinterpret_p16_s8
 
-poly16x4_t vreinterpret_p16_u64 (uint64x1_t t);
+_NEON2SSESTORAGE poly16x4_t vreinterpret_p16_u64 (uint64x1_t t);
 #define vreinterpret_p16_u64
 
-poly16x4_t vreinterpret_p16_s64 (int64x1_t t);
+_NEON2SSESTORAGE poly16x4_t vreinterpret_p16_s64 (int64x1_t t);
 #define vreinterpret_p16_s64
 
-poly16x4_t vreinterpret_p16_f32 (float32x2_t t);
+_NEON2SSESTORAGE poly16x4_t vreinterpret_p16_f32 (float32x2_t t);
 #define vreinterpret_p16_f32
 
-poly16x4_t vreinterpret_p16_p8 (poly8x8_t t);
+_NEON2SSESTORAGE poly16x4_t vreinterpret_p16_p8 (poly8x8_t t);
 #define vreinterpret_p16_p8
 
-poly16x8_t vreinterpretq_p16_u32 (uint32x4_t t);
+_NEON2SSESTORAGE poly16x8_t vreinterpretq_p16_u32 (uint32x4_t t);
 #define vreinterpretq_p16_u32
 
-poly16x8_t vreinterpretq_p16_u16 (uint16x8_t t);
+_NEON2SSESTORAGE poly16x8_t vreinterpretq_p16_u16 (uint16x8_t t);
 #define vreinterpretq_p16_u16
 
-poly16x8_t vreinterpretq_p16_s32 (int32x4_t t);
+_NEON2SSESTORAGE poly16x8_t vreinterpretq_p16_s32 (int32x4_t t);
 #define vreinterpretq_p16_s32
 
-poly16x8_t vreinterpretq_p16_s16 (int16x8_t t);
+_NEON2SSESTORAGE poly16x8_t vreinterpretq_p16_s16 (int16x8_t t);
 #define vreinterpretq_p16_s16
 
-poly16x8_t vreinterpretq_p16_s8 (int8x16_t t);
+_NEON2SSESTORAGE poly16x8_t vreinterpretq_p16_s8 (int8x16_t t);
 #define vreinterpretq_p16_s8
 
-poly16x8_t vreinterpretq_p16_u64 (uint64x2_t t);
+_NEON2SSESTORAGE poly16x8_t vreinterpretq_p16_u64 (uint64x2_t t);
 #define vreinterpretq_p16_u64
 
-poly16x8_t vreinterpretq_p16_s64 (int64x2_t t);
+_NEON2SSESTORAGE poly16x8_t vreinterpretq_p16_s64 (int64x2_t t);
 #define vreinterpretq_p16_s64
 
-poly16x8_t vreinterpretq_p16_f32 (float32x4_t t);
+_NEON2SSESTORAGE poly16x8_t vreinterpretq_p16_f32 (float32x4_t t);
 #define vreinterpretq_p16_f32(t) _M128i(t)
 
-poly16x8_t vreinterpretq_p16_p8 (poly8x16_t t);
+_NEON2SSESTORAGE poly16x8_t vreinterpretq_p16_p8 (poly8x16_t t);
 #define vreinterpretq_p16_p8  vreinterpretq_s16_p8
 
 //****  Integer to float  ******
-float32x2_t vreinterpret_f32_u32 (uint32x2_t t);
+_NEON2SSESTORAGE float32x2_t vreinterpret_f32_u32 (uint32x2_t t);
 _NEON2SSE_INLINE float32x2_t vreinterpret_f32_u32 (uint32x2_t t)
 {
     return (*(__m64_128*)&(t));
 }
 
-float32x2_t vreinterpret_f32_u16 (uint16x4_t t);
+_NEON2SSESTORAGE float32x2_t vreinterpret_f32_u16 (uint16x4_t t);
 #define vreinterpret_f32_u16 vreinterpret_f32_u32
 
 
-float32x2_t vreinterpret_f32_u8 (uint8x8_t t);
+_NEON2SSESTORAGE float32x2_t vreinterpret_f32_u8 (uint8x8_t t);
 #define vreinterpret_f32_u8 vreinterpret_f32_u32
 
 
-float32x2_t vreinterpret_f32_s32 (int32x2_t t);
+_NEON2SSESTORAGE float32x2_t vreinterpret_f32_s32 (int32x2_t t);
 #define vreinterpret_f32_s32 vreinterpret_f32_u32
 
 
-float32x2_t vreinterpret_f32_s16 (int16x4_t t);
+_NEON2SSESTORAGE float32x2_t vreinterpret_f32_s16 (int16x4_t t);
 #define vreinterpret_f32_s16 vreinterpret_f32_u32
 
-float32x2_t vreinterpret_f32_s8 (int8x8_t t);
+_NEON2SSESTORAGE float32x2_t vreinterpret_f32_s8 (int8x8_t t);
 #define vreinterpret_f32_s8 vreinterpret_f32_u32
 
 
-float32x2_t vreinterpret_f32_u64(uint64x1_t t);
+_NEON2SSESTORAGE float32x2_t vreinterpret_f32_u64(uint64x1_t t);
 #define vreinterpret_f32_u64 vreinterpret_f32_u32
 
 
-float32x2_t vreinterpret_f32_s64 (int64x1_t t);
+_NEON2SSESTORAGE float32x2_t vreinterpret_f32_s64 (int64x1_t t);
 #define vreinterpret_f32_s64 vreinterpret_f32_u32
 
 
-float32x2_t vreinterpret_f32_p16 (poly16x4_t t);
+_NEON2SSESTORAGE float32x2_t vreinterpret_f32_p16 (poly16x4_t t);
 #define vreinterpret_f32_p16 vreinterpret_f32_u32
 
-float32x2_t vreinterpret_f32_p8 (poly8x8_t t);
+_NEON2SSESTORAGE float32x2_t vreinterpret_f32_p8 (poly8x8_t t);
 #define vreinterpret_f32_p8 vreinterpret_f32_u32
 
-float32x4_t vreinterpretq_f32_u32 (uint32x4_t t);
+_NEON2SSESTORAGE float32x4_t vreinterpretq_f32_u32 (uint32x4_t t);
 #define  vreinterpretq_f32_u32(t) _M128(t)
 
-float32x4_t vreinterpretq_f32_u16 (uint16x8_t t);
+_NEON2SSESTORAGE float32x4_t vreinterpretq_f32_u16 (uint16x8_t t);
 #define vreinterpretq_f32_u16 vreinterpretq_f32_u32
 
-float32x4_t vreinterpretq_f32_u8 (uint8x16_t t);
+_NEON2SSESTORAGE float32x4_t vreinterpretq_f32_u8 (uint8x16_t t);
 #define vreinterpretq_f32_u8 vreinterpretq_f32_u32
 
-float32x4_t vreinterpretq_f32_s32 (int32x4_t t);
+_NEON2SSESTORAGE float32x4_t vreinterpretq_f32_s32 (int32x4_t t);
 #define vreinterpretq_f32_s32 vreinterpretq_f32_u32
 
-float32x4_t vreinterpretq_f32_s16 (int16x8_t t);
+_NEON2SSESTORAGE float32x4_t vreinterpretq_f32_s16 (int16x8_t t);
 #define vreinterpretq_f32_s16 vreinterpretq_f32_u32
 
-float32x4_t vreinterpretq_f32_s8 (int8x16_t t);
+_NEON2SSESTORAGE float32x4_t vreinterpretq_f32_s8 (int8x16_t t);
 #define vreinterpretq_f32_s8 vreinterpretq_f32_u32
 
-float32x4_t vreinterpretq_f32_u64 (uint64x2_t t);
+_NEON2SSESTORAGE float32x4_t vreinterpretq_f32_u64 (uint64x2_t t);
 #define vreinterpretq_f32_u64 vreinterpretq_f32_u32
 
-float32x4_t vreinterpretq_f32_s64 (int64x2_t t);
+_NEON2SSESTORAGE float32x4_t vreinterpretq_f32_s64 (int64x2_t t);
 #define vreinterpretq_f32_s64 vreinterpretq_f32_u32
 
-float32x4_t vreinterpretq_f32_p16 (poly16x8_t t);
+_NEON2SSESTORAGE float32x4_t vreinterpretq_f32_p16 (poly16x8_t t);
 #define vreinterpretq_f32_p16 vreinterpretq_f32_u32
 
-float32x4_t vreinterpretq_f32_p8 (poly8x16_t t);
+_NEON2SSESTORAGE float32x4_t vreinterpretq_f32_p8 (poly8x16_t t);
 #define vreinterpretq_f32_p8 vreinterpretq_f32_u32
 
 //*** Integer type conversions ******************
 //no conversion necessary for the following functions because it is same data type
-int64x1_t vreinterpret_s64_u32 (uint32x2_t t);
+_NEON2SSESTORAGE int64x1_t vreinterpret_s64_u32 (uint32x2_t t);
 #define vreinterpret_s64_u32
 
-int64x1_t vreinterpret_s64_u16 (uint16x4_t t);
+_NEON2SSESTORAGE int64x1_t vreinterpret_s64_u16 (uint16x4_t t);
 #define vreinterpret_s64_u16
 
-int64x1_t vreinterpret_s64_u8 (uint8x8_t t);
+_NEON2SSESTORAGE int64x1_t vreinterpret_s64_u8 (uint8x8_t t);
 #define vreinterpret_s64_u8
 
-int64x1_t vreinterpret_s64_s32 (int32x2_t t);
+_NEON2SSESTORAGE int64x1_t vreinterpret_s64_s32 (int32x2_t t);
 #define  vreinterpret_s64_s32
 
-int64x1_t vreinterpret_s64_s16 (int16x4_t t);
+_NEON2SSESTORAGE int64x1_t vreinterpret_s64_s16 (int16x4_t t);
 #define vreinterpret_s64_s16
 
-int64x1_t vreinterpret_s64_s8 (int8x8_t t);
+_NEON2SSESTORAGE int64x1_t vreinterpret_s64_s8 (int8x8_t t);
 #define  vreinterpret_s64_s8
 
-int64x1_t vreinterpret_s64_u64 (uint64x1_t t);
+_NEON2SSESTORAGE int64x1_t vreinterpret_s64_u64 (uint64x1_t t);
 #define  vreinterpret_s64_u64
 
-int64x1_t vreinterpret_s64_f32 (float32x2_t t);
+_NEON2SSESTORAGE int64x1_t vreinterpret_s64_f32 (float32x2_t t);
 #define  vreinterpret_s64_f32
 
-int64x1_t vreinterpret_s64_p16 (poly16x4_t t);
+_NEON2SSESTORAGE int64x1_t vreinterpret_s64_p16 (poly16x4_t t);
 #define vreinterpret_s64_p16
 
-int64x1_t vreinterpret_s64_p8 (poly8x8_t t);
+_NEON2SSESTORAGE int64x1_t vreinterpret_s64_p8 (poly8x8_t t);
 #define vreinterpret_s64_p8
 
-int64x2_t vreinterpretq_s64_u32 (uint32x4_t t);
+_NEON2SSESTORAGE int64x2_t vreinterpretq_s64_u32 (uint32x4_t t);
 #define vreinterpretq_s64_u32
 
-int64x2_t vreinterpretq_s64_s16 (uint16x8_t t);
+_NEON2SSESTORAGE int64x2_t vreinterpretq_s64_s16 (uint16x8_t t);
 #define vreinterpretq_s64_s16
 
-int64x2_t vreinterpretq_s64_u8 (uint8x16_t t);
+_NEON2SSESTORAGE int64x2_t vreinterpretq_s64_u8 (uint8x16_t t);
 #define vreinterpretq_s64_u8
 
-int64x2_t vreinterpretq_s64_s32 (int32x4_t t);
+_NEON2SSESTORAGE int64x2_t vreinterpretq_s64_s32 (int32x4_t t);
 #define vreinterpretq_s64_s32
 
-int64x2_t vreinterpretq_s64_u16 (int16x8_t t);
+_NEON2SSESTORAGE int64x2_t vreinterpretq_s64_u16 (int16x8_t t);
 #define vreinterpretq_s64_u16
 
-int64x2_t vreinterpretq_s64_s8 (int8x16_t t);
+_NEON2SSESTORAGE int64x2_t vreinterpretq_s64_s8 (int8x16_t t);
 #define vreinterpretq_s64_s8
 
-int64x2_t vreinterpretq_s64_u64 (uint64x2_t t);
+_NEON2SSESTORAGE int64x2_t vreinterpretq_s64_u64 (uint64x2_t t);
 #define vreinterpretq_s64_u64
 
-int64x2_t vreinterpretq_s64_f32 (float32x4_t t);
+_NEON2SSESTORAGE int64x2_t vreinterpretq_s64_f32 (float32x4_t t);
 #define vreinterpretq_s64_f32(t) _M128i(t)
 
-int64x2_t vreinterpretq_s64_p16 (poly16x8_t t);
+_NEON2SSESTORAGE int64x2_t vreinterpretq_s64_p16 (poly16x8_t t);
 #define vreinterpretq_s64_p16
 
-int64x2_t vreinterpretq_s64_p8 (poly8x16_t t);
+_NEON2SSESTORAGE int64x2_t vreinterpretq_s64_p8 (poly8x16_t t);
 #define vreinterpretq_s64_p8
 
-uint64x1_t vreinterpret_u64_u32 (uint32x2_t t);
+_NEON2SSESTORAGE uint64x1_t vreinterpret_u64_u32 (uint32x2_t t);
 #define vreinterpret_u64_u32
 
-uint64x1_t vreinterpret_u64_u16 (uint16x4_t t);
+_NEON2SSESTORAGE uint64x1_t vreinterpret_u64_u16 (uint16x4_t t);
 #define vreinterpret_u64_u16
 
-uint64x1_t vreinterpret_u64_u8 (uint8x8_t t);
+_NEON2SSESTORAGE uint64x1_t vreinterpret_u64_u8 (uint8x8_t t);
 #define vreinterpret_u64_u8
 
-uint64x1_t vreinterpret_u64_s32 (int32x2_t t);
+_NEON2SSESTORAGE uint64x1_t vreinterpret_u64_s32 (int32x2_t t);
 #define vreinterpret_u64_s32
 
-uint64x1_t vreinterpret_u64_s16 (int16x4_t t);
+_NEON2SSESTORAGE uint64x1_t vreinterpret_u64_s16 (int16x4_t t);
 #define vreinterpret_u64_s16
 
-uint64x1_t vreinterpret_u64_s8 (int8x8_t t);
+_NEON2SSESTORAGE uint64x1_t vreinterpret_u64_s8 (int8x8_t t);
 #define vreinterpret_u64_s8
 
-uint64x1_t vreinterpret_u64_s64 (int64x1_t t);
+_NEON2SSESTORAGE uint64x1_t vreinterpret_u64_s64 (int64x1_t t);
 #define vreinterpret_u64_s64
 
-uint64x1_t vreinterpret_u64_f32 (float32x2_t t);
+_NEON2SSESTORAGE uint64x1_t vreinterpret_u64_f32 (float32x2_t t);
 #define vreinterpret_u64_f32
 
-uint64x1_t vreinterpret_u64_p16 (poly16x4_t t);
+_NEON2SSESTORAGE uint64x1_t vreinterpret_u64_p16 (poly16x4_t t);
 #define vreinterpret_u64_p16
 
-uint64x1_t vreinterpret_u64_p8 (poly8x8_t t);
+_NEON2SSESTORAGE uint64x1_t vreinterpret_u64_p8 (poly8x8_t t);
 #define vreinterpret_u64_p8
 
-uint64x2_t vreinterpretq_u64_u32 (uint32x4_t t);
+_NEON2SSESTORAGE uint64x2_t vreinterpretq_u64_u32 (uint32x4_t t);
 #define vreinterpretq_u64_u32
 
-uint64x2_t vreinterpretq_u64_u16 (uint16x8_t t);
+_NEON2SSESTORAGE uint64x2_t vreinterpretq_u64_u16 (uint16x8_t t);
 #define vreinterpretq_u64_u16
 
-uint64x2_t vreinterpretq_u64_u8 (uint8x16_t t);
+_NEON2SSESTORAGE uint64x2_t vreinterpretq_u64_u8 (uint8x16_t t);
 #define vreinterpretq_u64_u8
 
-uint64x2_t vreinterpretq_u64_s32 (int32x4_t t);
+_NEON2SSESTORAGE uint64x2_t vreinterpretq_u64_s32 (int32x4_t t);
 #define vreinterpretq_u64_s32
 
-uint64x2_t vreinterpretq_u64_s16 (int16x8_t t);
+_NEON2SSESTORAGE uint64x2_t vreinterpretq_u64_s16 (int16x8_t t);
 #define vreinterpretq_u64_s16
 
-uint64x2_t vreinterpretq_u64_s8 (int8x16_t t);
+_NEON2SSESTORAGE uint64x2_t vreinterpretq_u64_s8 (int8x16_t t);
 #define vreinterpretq_u64_s8
 
-uint64x2_t vreinterpretq_u64_s64 (int64x2_t t);
+_NEON2SSESTORAGE uint64x2_t vreinterpretq_u64_s64 (int64x2_t t);
 #define vreinterpretq_u64_s64
 
-uint64x2_t vreinterpretq_u64_f32 (float32x4_t t);
+_NEON2SSESTORAGE uint64x2_t vreinterpretq_u64_f32 (float32x4_t t);
 #define vreinterpretq_u64_f32(t) _M128i(t)
 
-uint64x2_t vreinterpretq_u64_p16 (poly16x8_t t);
+_NEON2SSESTORAGE uint64x2_t vreinterpretq_u64_p16 (poly16x8_t t);
 #define vreinterpretq_u64_p16
 
-uint64x2_t vreinterpretq_u64_p8 (poly8x16_t t);
+_NEON2SSESTORAGE uint64x2_t vreinterpretq_u64_p8 (poly8x16_t t);
 #define vreinterpretq_u64_p8
 
-int8x8_t vreinterpret_s8_u32 (uint32x2_t t);
+_NEON2SSESTORAGE int8x8_t vreinterpret_s8_u32 (uint32x2_t t);
 #define vreinterpret_s8_u32
 
-int8x8_t vreinterpret_s8_u16 (uint16x4_t t);
+_NEON2SSESTORAGE int8x8_t vreinterpret_s8_u16 (uint16x4_t t);
 #define vreinterpret_s8_u16
 
-int8x8_t vreinterpret_s8_u8 (uint8x8_t t);
+_NEON2SSESTORAGE int8x8_t vreinterpret_s8_u8 (uint8x8_t t);
 #define vreinterpret_s8_u8
 
-int8x8_t vreinterpret_s8_s32 (int32x2_t t);
+_NEON2SSESTORAGE int8x8_t vreinterpret_s8_s32 (int32x2_t t);
 #define vreinterpret_s8_s32
 
-int8x8_t vreinterpret_s8_s16 (int16x4_t t);
+_NEON2SSESTORAGE int8x8_t vreinterpret_s8_s16 (int16x4_t t);
 #define vreinterpret_s8_s16
 
-int8x8_t vreinterpret_s8_u64 (uint64x1_t t);
+_NEON2SSESTORAGE int8x8_t vreinterpret_s8_u64 (uint64x1_t t);
 #define vreinterpret_s8_u64
 
-int8x8_t vreinterpret_s8_s64 (int64x1_t t);
+_NEON2SSESTORAGE int8x8_t vreinterpret_s8_s64 (int64x1_t t);
 #define vreinterpret_s8_s64
 
-int8x8_t vreinterpret_s8_f32 (float32x2_t t);
+_NEON2SSESTORAGE int8x8_t vreinterpret_s8_f32 (float32x2_t t);
 #define vreinterpret_s8_f32
 
-int8x8_t vreinterpret_s8_p16 (poly16x4_t t);
+_NEON2SSESTORAGE int8x8_t vreinterpret_s8_p16 (poly16x4_t t);
 #define vreinterpret_s8_p16
 
-int8x8_t vreinterpret_s8_p8 (poly8x8_t t);
+_NEON2SSESTORAGE int8x8_t vreinterpret_s8_p8 (poly8x8_t t);
 #define vreinterpret_s8_p8
 
-int8x16_t vreinterpretq_s8_u32 (uint32x4_t t);
+_NEON2SSESTORAGE int8x16_t vreinterpretq_s8_u32 (uint32x4_t t);
 #define vreinterpretq_s8_u32
 
-int8x16_t vreinterpretq_s8_u16 (uint16x8_t t);
+_NEON2SSESTORAGE int8x16_t vreinterpretq_s8_u16 (uint16x8_t t);
 #define vreinterpretq_s8_u16
 
-int8x16_t vreinterpretq_s8_u8 (uint8x16_t t);
+_NEON2SSESTORAGE int8x16_t vreinterpretq_s8_u8 (uint8x16_t t);
 #define vreinterpretq_s8_u8
 
-int8x16_t vreinterpretq_s8_s32 (int32x4_t t);
+_NEON2SSESTORAGE int8x16_t vreinterpretq_s8_s32 (int32x4_t t);
 #define vreinterpretq_s8_s32
 
-int8x16_t vreinterpretq_s8_s16 (int16x8_t t);
+_NEON2SSESTORAGE int8x16_t vreinterpretq_s8_s16 (int16x8_t t);
 #define vreinterpretq_s8_s16
 
-int8x16_t vreinterpretq_s8_u64 (uint64x2_t t);
+_NEON2SSESTORAGE int8x16_t vreinterpretq_s8_u64 (uint64x2_t t);
 #define vreinterpretq_s8_u64
 
-int8x16_t vreinterpretq_s8_s64 (int64x2_t t);
+_NEON2SSESTORAGE int8x16_t vreinterpretq_s8_s64 (int64x2_t t);
 #define vreinterpretq_s8_s64
 
-int8x16_t vreinterpretq_s8_f32 (float32x4_t t);
+_NEON2SSESTORAGE int8x16_t vreinterpretq_s8_f32 (float32x4_t t);
 #define vreinterpretq_s8_f32(t) _M128i(t)
 
-int8x16_t vreinterpretq_s8_p16 (poly16x8_t t);
+_NEON2SSESTORAGE int8x16_t vreinterpretq_s8_p16 (poly16x8_t t);
 #define vreinterpretq_s8_p16
 
-int8x16_t vreinterpretq_s8_p8 (poly8x16_t t);
+_NEON2SSESTORAGE int8x16_t vreinterpretq_s8_p8 (poly8x16_t t);
 #define vreinterpretq_s8_p8
 
-int16x4_t vreinterpret_s16_u32 (uint32x2_t t);
+_NEON2SSESTORAGE int16x4_t vreinterpret_s16_u32 (uint32x2_t t);
 #define vreinterpret_s16_u32
 
-int16x4_t vreinterpret_s16_u16 (uint16x4_t t);
+_NEON2SSESTORAGE int16x4_t vreinterpret_s16_u16 (uint16x4_t t);
 #define vreinterpret_s16_u16
 
-int16x4_t vreinterpret_s16_u8 (uint8x8_t t);
+_NEON2SSESTORAGE int16x4_t vreinterpret_s16_u8 (uint8x8_t t);
 #define vreinterpret_s16_u8
 
-int16x4_t vreinterpret_s16_s32 (int32x2_t t);
+_NEON2SSESTORAGE int16x4_t vreinterpret_s16_s32 (int32x2_t t);
 #define vreinterpret_s16_s32
 
-int16x4_t vreinterpret_s16_s8 (int8x8_t t);
+_NEON2SSESTORAGE int16x4_t vreinterpret_s16_s8 (int8x8_t t);
 #define vreinterpret_s16_s8
 
-int16x4_t vreinterpret_s16_u64 (uint64x1_t t);
+_NEON2SSESTORAGE int16x4_t vreinterpret_s16_u64 (uint64x1_t t);
 #define vreinterpret_s16_u64
 
-int16x4_t vreinterpret_s16_s64 (int64x1_t t);
+_NEON2SSESTORAGE int16x4_t vreinterpret_s16_s64 (int64x1_t t);
 #define vreinterpret_s16_s64
 
-int16x4_t vreinterpret_s16_f32 (float32x2_t t);
+_NEON2SSESTORAGE int16x4_t vreinterpret_s16_f32 (float32x2_t t);
 #define vreinterpret_s16_f32
 
 
-int16x4_t vreinterpret_s16_p16 (poly16x4_t t);
+_NEON2SSESTORAGE int16x4_t vreinterpret_s16_p16 (poly16x4_t t);
 #define vreinterpret_s16_p16
 
-int16x4_t vreinterpret_s16_p8 (poly8x8_t t);
+_NEON2SSESTORAGE int16x4_t vreinterpret_s16_p8 (poly8x8_t t);
 #define vreinterpret_s16_p8
 
-int16x8_t vreinterpretq_s16_u32 (uint32x4_t t);
+_NEON2SSESTORAGE int16x8_t vreinterpretq_s16_u32 (uint32x4_t t);
 #define vreinterpretq_s16_u32
 
-int16x8_t vreinterpretq_s16_u16 (uint16x8_t t);
+_NEON2SSESTORAGE int16x8_t vreinterpretq_s16_u16 (uint16x8_t t);
 #define vreinterpretq_s16_u16
 
-int16x8_t vreinterpretq_s16_u8 (uint8x16_t t);
+_NEON2SSESTORAGE int16x8_t vreinterpretq_s16_u8 (uint8x16_t t);
 #define vreinterpretq_s16_u8
 
-int16x8_t vreinterpretq_s16_s32 (int32x4_t t);
+_NEON2SSESTORAGE int16x8_t vreinterpretq_s16_s32 (int32x4_t t);
 #define vreinterpretq_s16_s32
 
-int16x8_t vreinterpretq_s16_s8 (int8x16_t t);
+_NEON2SSESTORAGE int16x8_t vreinterpretq_s16_s8 (int8x16_t t);
 #define vreinterpretq_s16_s8
 
-int16x8_t vreinterpretq_s16_u64 (uint64x2_t t);
+_NEON2SSESTORAGE int16x8_t vreinterpretq_s16_u64 (uint64x2_t t);
 #define vreinterpretq_s16_u64
 
-int16x8_t vreinterpretq_s16_s64 (int64x2_t t);
+_NEON2SSESTORAGE int16x8_t vreinterpretq_s16_s64 (int64x2_t t);
 #define vreinterpretq_s16_s64
 
-int16x8_t vreinterpretq_s16_f32 (float32x4_t t);
+_NEON2SSESTORAGE int16x8_t vreinterpretq_s16_f32 (float32x4_t t);
 #define vreinterpretq_s16_f32(t) _M128i(t)
 
-int16x8_t vreinterpretq_s16_p16 (poly16x8_t t);
+_NEON2SSESTORAGE int16x8_t vreinterpretq_s16_p16 (poly16x8_t t);
 #define vreinterpretq_s16_p16
 
-int16x8_t vreinterpretq_s16_p8 (poly8x16_t t);
+_NEON2SSESTORAGE int16x8_t vreinterpretq_s16_p8 (poly8x16_t t);
 #define vreinterpretq_s16_p8
 
-int32x2_t vreinterpret_s32_u32 (uint32x2_t t);
+_NEON2SSESTORAGE int32x2_t vreinterpret_s32_u32 (uint32x2_t t);
 #define vreinterpret_s32_u32
 
-int32x2_t vreinterpret_s32_u16 (uint16x4_t t);
+_NEON2SSESTORAGE int32x2_t vreinterpret_s32_u16 (uint16x4_t t);
 #define vreinterpret_s32_u16
 
-int32x2_t vreinterpret_s32_u8 (uint8x8_t t);
+_NEON2SSESTORAGE int32x2_t vreinterpret_s32_u8 (uint8x8_t t);
 #define vreinterpret_s32_u8
 
-int32x2_t vreinterpret_s32_s16 (int16x4_t t);
+_NEON2SSESTORAGE int32x2_t vreinterpret_s32_s16 (int16x4_t t);
 #define vreinterpret_s32_s16
 
-int32x2_t vreinterpret_s32_s8 (int8x8_t t);
+_NEON2SSESTORAGE int32x2_t vreinterpret_s32_s8 (int8x8_t t);
 #define vreinterpret_s32_s8
 
-int32x2_t vreinterpret_s32_u64 (uint64x1_t t);
+_NEON2SSESTORAGE int32x2_t vreinterpret_s32_u64 (uint64x1_t t);
 #define vreinterpret_s32_u64
 
-int32x2_t vreinterpret_s32_s64 (int64x1_t t);
+_NEON2SSESTORAGE int32x2_t vreinterpret_s32_s64 (int64x1_t t);
 #define vreinterpret_s32_s64
 
-int32x2_t vreinterpret_s32_f32 (float32x2_t t);
+_NEON2SSESTORAGE int32x2_t vreinterpret_s32_f32 (float32x2_t t);
 #define vreinterpret_s32_f32
 
-int32x2_t vreinterpret_s32_p16 (poly16x4_t t);
+_NEON2SSESTORAGE int32x2_t vreinterpret_s32_p16 (poly16x4_t t);
 #define vreinterpret_s32_p16
 
-int32x2_t vreinterpret_s32_p8 (poly8x8_t t);
+_NEON2SSESTORAGE int32x2_t vreinterpret_s32_p8 (poly8x8_t t);
 #define vreinterpret_s32_p8
 
-int32x4_t vreinterpretq_s32_u32 (uint32x4_t t);
+_NEON2SSESTORAGE int32x4_t vreinterpretq_s32_u32 (uint32x4_t t);
 #define vreinterpretq_s32_u32
 
-int32x4_t vreinterpretq_s32_u16 (uint16x8_t t);
+_NEON2SSESTORAGE int32x4_t vreinterpretq_s32_u16 (uint16x8_t t);
 #define vreinterpretq_s32_u16
 
-int32x4_t vreinterpretq_s32_u8 (uint8x16_t t);
+_NEON2SSESTORAGE int32x4_t vreinterpretq_s32_u8 (uint8x16_t t);
 #define vreinterpretq_s32_u8
 
-int32x4_t vreinterpretq_s32_s16 (int16x8_t t);
+_NEON2SSESTORAGE int32x4_t vreinterpretq_s32_s16 (int16x8_t t);
 #define vreinterpretq_s32_s16
 
-int32x4_t vreinterpretq_s32_s8 (int8x16_t t);
+_NEON2SSESTORAGE int32x4_t vreinterpretq_s32_s8 (int8x16_t t);
 #define vreinterpretq_s32_s8
 
-int32x4_t vreinterpretq_s32_u64 (uint64x2_t t);
+_NEON2SSESTORAGE int32x4_t vreinterpretq_s32_u64 (uint64x2_t t);
 #define vreinterpretq_s32_u64
 
-int32x4_t vreinterpretq_s32_s64 (int64x2_t t);
+_NEON2SSESTORAGE int32x4_t vreinterpretq_s32_s64 (int64x2_t t);
 #define vreinterpretq_s32_s64
 
-int32x4_t vreinterpretq_s32_f32 (float32x4_t t);
-#define vreinterpretq_s32_f32(t)  _M128i(t) 
+_NEON2SSESTORAGE int32x4_t vreinterpretq_s32_f32 (float32x4_t t);
+#define vreinterpretq_s32_f32(t)  _M128i(t)
 
-int32x4_t vreinterpretq_s32_p16 (poly16x8_t t);
+_NEON2SSESTORAGE int32x4_t vreinterpretq_s32_p16 (poly16x8_t t);
 #define vreinterpretq_s32_p16
 
-int32x4_t vreinterpretq_s32_p8 (poly8x16_t t);
+_NEON2SSESTORAGE int32x4_t vreinterpretq_s32_p8 (poly8x16_t t);
 #define vreinterpretq_s32_p8
 
-uint8x8_t vreinterpret_u8_u32 (uint32x2_t t);
+_NEON2SSESTORAGE uint8x8_t vreinterpret_u8_u32 (uint32x2_t t);
 #define vreinterpret_u8_u32
 
-uint8x8_t vreinterpret_u8_u16 (uint16x4_t t);
+_NEON2SSESTORAGE uint8x8_t vreinterpret_u8_u16 (uint16x4_t t);
 #define vreinterpret_u8_u16
 
-uint8x8_t vreinterpret_u8_s32 (int32x2_t t);
+_NEON2SSESTORAGE uint8x8_t vreinterpret_u8_s32 (int32x2_t t);
 #define vreinterpret_u8_s32
 
-uint8x8_t vreinterpret_u8_s16 (int16x4_t t);
+_NEON2SSESTORAGE uint8x8_t vreinterpret_u8_s16 (int16x4_t t);
 #define vreinterpret_u8_s16
 
-uint8x8_t vreinterpret_u8_s8 (int8x8_t t);
+_NEON2SSESTORAGE uint8x8_t vreinterpret_u8_s8 (int8x8_t t);
 #define vreinterpret_u8_s8
 
-uint8x8_t vreinterpret_u8_u64 (uint64x1_t t);
+_NEON2SSESTORAGE uint8x8_t vreinterpret_u8_u64 (uint64x1_t t);
 #define vreinterpret_u8_u64
 
-uint8x8_t vreinterpret_u8_s64 (int64x1_t t);
+_NEON2SSESTORAGE uint8x8_t vreinterpret_u8_s64 (int64x1_t t);
 #define vreinterpret_u8_s64
 
-uint8x8_t vreinterpret_u8_f32 (float32x2_t t);
+_NEON2SSESTORAGE uint8x8_t vreinterpret_u8_f32 (float32x2_t t);
 #define vreinterpret_u8_f32
 
-uint8x8_t vreinterpret_u8_p16 (poly16x4_t t);
+_NEON2SSESTORAGE uint8x8_t vreinterpret_u8_p16 (poly16x4_t t);
 #define vreinterpret_u8_p16
 
-uint8x8_t vreinterpret_u8_p8 (poly8x8_t t);
+_NEON2SSESTORAGE uint8x8_t vreinterpret_u8_p8 (poly8x8_t t);
 #define vreinterpret_u8_p8
 
-uint8x16_t vreinterpretq_u8_u32 (uint32x4_t t);
+_NEON2SSESTORAGE uint8x16_t vreinterpretq_u8_u32 (uint32x4_t t);
 #define vreinterpretq_u8_u32
 
-uint8x16_t vreinterpretq_u8_u16 (uint16x8_t t);
+_NEON2SSESTORAGE uint8x16_t vreinterpretq_u8_u16 (uint16x8_t t);
 #define vreinterpretq_u8_u16
 
-uint8x16_t vreinterpretq_u8_s32 (int32x4_t t);
+_NEON2SSESTORAGE uint8x16_t vreinterpretq_u8_s32 (int32x4_t t);
 #define vreinterpretq_u8_s32
 
-uint8x16_t vreinterpretq_u8_s16 (int16x8_t t);
+_NEON2SSESTORAGE uint8x16_t vreinterpretq_u8_s16 (int16x8_t t);
 #define vreinterpretq_u8_s16
 
-uint8x16_t vreinterpretq_u8_s8 (int8x16_t t);
+_NEON2SSESTORAGE uint8x16_t vreinterpretq_u8_s8 (int8x16_t t);
 #define vreinterpretq_u8_s8
 
-uint8x16_t vreinterpretq_u8_u64 (uint64x2_t t);
+_NEON2SSESTORAGE uint8x16_t vreinterpretq_u8_u64 (uint64x2_t t);
 #define vreinterpretq_u8_u64
 
-uint8x16_t vreinterpretq_u8_s64 (int64x2_t t);
+_NEON2SSESTORAGE uint8x16_t vreinterpretq_u8_s64 (int64x2_t t);
 #define vreinterpretq_u8_s64
 
-uint8x16_t vreinterpretq_u8_f32 (float32x4_t t);
+_NEON2SSESTORAGE uint8x16_t vreinterpretq_u8_f32 (float32x4_t t);
 #define vreinterpretq_u8_f32(t) _M128i(t)
 
 
-uint8x16_t vreinterpretq_u8_p16 (poly16x8_t t);
+_NEON2SSESTORAGE uint8x16_t vreinterpretq_u8_p16 (poly16x8_t t);
 #define vreinterpretq_u8_p16
 
-uint8x16_t vreinterpretq_u8_p8 (poly8x16_t t);
+_NEON2SSESTORAGE uint8x16_t vreinterpretq_u8_p8 (poly8x16_t t);
 #define vreinterpretq_u8_p8
 
-uint16x4_t vreinterpret_u16_u32 (uint32x2_t t);
+_NEON2SSESTORAGE uint16x4_t vreinterpret_u16_u32 (uint32x2_t t);
 #define vreinterpret_u16_u32
 
-uint16x4_t vreinterpret_u16_u8 (uint8x8_t t);
+_NEON2SSESTORAGE uint16x4_t vreinterpret_u16_u8 (uint8x8_t t);
 #define vreinterpret_u16_u8
 
-uint16x4_t vreinterpret_u16_s32 (int32x2_t t);
+_NEON2SSESTORAGE uint16x4_t vreinterpret_u16_s32 (int32x2_t t);
 #define vreinterpret_u16_s32
 
-uint16x4_t vreinterpret_u16_s16 (int16x4_t t);
+_NEON2SSESTORAGE uint16x4_t vreinterpret_u16_s16 (int16x4_t t);
 #define vreinterpret_u16_s16
 
-uint16x4_t vreinterpret_u16_s8 (int8x8_t t);
+_NEON2SSESTORAGE uint16x4_t vreinterpret_u16_s8 (int8x8_t t);
 #define vreinterpret_u16_s8
 
-uint16x4_t vreinterpret_u16_u64 (uint64x1_t t);
+_NEON2SSESTORAGE uint16x4_t vreinterpret_u16_u64 (uint64x1_t t);
 #define vreinterpret_u16_u64
 
-uint16x4_t vreinterpret_u16_s64 (int64x1_t t);
+_NEON2SSESTORAGE uint16x4_t vreinterpret_u16_s64 (int64x1_t t);
 #define vreinterpret_u16_s64
 
-uint16x4_t vreinterpret_u16_f32 (float32x2_t t);
+_NEON2SSESTORAGE uint16x4_t vreinterpret_u16_f32 (float32x2_t t);
 #define vreinterpret_u16_f32
 
-uint16x4_t vreinterpret_u16_p16 (poly16x4_t t);
+_NEON2SSESTORAGE uint16x4_t vreinterpret_u16_p16 (poly16x4_t t);
 #define vreinterpret_u16_p16
 
-uint16x4_t vreinterpret_u16_p8 (poly8x8_t t);
+_NEON2SSESTORAGE uint16x4_t vreinterpret_u16_p8 (poly8x8_t t);
 #define vreinterpret_u16_p8
 
-uint16x8_t vreinterpretq_u16_u32 (uint32x4_t t);
+_NEON2SSESTORAGE uint16x8_t vreinterpretq_u16_u32 (uint32x4_t t);
 #define vreinterpretq_u16_u32
 
-uint16x8_t vreinterpretq_u16_u8 (uint8x16_t t);
+_NEON2SSESTORAGE uint16x8_t vreinterpretq_u16_u8 (uint8x16_t t);
 #define vreinterpretq_u16_u8
 
-uint16x8_t vreinterpretq_u16_s32 (int32x4_t t);
+_NEON2SSESTORAGE uint16x8_t vreinterpretq_u16_s32 (int32x4_t t);
 #define vreinterpretq_u16_s32
 
-uint16x8_t vreinterpretq_u16_s16 (int16x8_t t);
+_NEON2SSESTORAGE uint16x8_t vreinterpretq_u16_s16 (int16x8_t t);
 #define vreinterpretq_u16_s16
 
-uint16x8_t vreinterpretq_u16_s8 (int8x16_t t);
+_NEON2SSESTORAGE uint16x8_t vreinterpretq_u16_s8 (int8x16_t t);
 #define vreinterpretq_u16_s8
 
-uint16x8_t vreinterpretq_u16_u64 (uint64x2_t t);
+_NEON2SSESTORAGE uint16x8_t vreinterpretq_u16_u64 (uint64x2_t t);
 #define vreinterpretq_u16_u64
 
-uint16x8_t vreinterpretq_u16_s64 (int64x2_t t);
+_NEON2SSESTORAGE uint16x8_t vreinterpretq_u16_s64 (int64x2_t t);
 #define vreinterpretq_u16_s64
 
-uint16x8_t vreinterpretq_u16_f32 (float32x4_t t);
+_NEON2SSESTORAGE uint16x8_t vreinterpretq_u16_f32 (float32x4_t t);
 #define vreinterpretq_u16_f32(t) _M128i(t)
 
-uint16x8_t vreinterpretq_u16_p16 (poly16x8_t t);
+_NEON2SSESTORAGE uint16x8_t vreinterpretq_u16_p16 (poly16x8_t t);
 #define vreinterpretq_u16_p16
 
-uint16x8_t vreinterpretq_u16_p8 (poly8x16_t t);
+_NEON2SSESTORAGE uint16x8_t vreinterpretq_u16_p8 (poly8x16_t t);
 #define vreinterpretq_u16_p8
 
-uint32x2_t vreinterpret_u32_u16 (uint16x4_t t);
+_NEON2SSESTORAGE uint32x2_t vreinterpret_u32_u16 (uint16x4_t t);
 #define vreinterpret_u32_u16
 
-uint32x2_t vreinterpret_u32_u8 (uint8x8_t t);
+_NEON2SSESTORAGE uint32x2_t vreinterpret_u32_u8 (uint8x8_t t);
 #define vreinterpret_u32_u8
 
-uint32x2_t vreinterpret_u32_s32 (int32x2_t t);
+_NEON2SSESTORAGE uint32x2_t vreinterpret_u32_s32 (int32x2_t t);
 #define vreinterpret_u32_s32
 
-uint32x2_t vreinterpret_u32_s16 (int16x4_t t);
+_NEON2SSESTORAGE uint32x2_t vreinterpret_u32_s16 (int16x4_t t);
 #define vreinterpret_u32_s16
 
-uint32x2_t vreinterpret_u32_s8 (int8x8_t t);
+_NEON2SSESTORAGE uint32x2_t vreinterpret_u32_s8 (int8x8_t t);
 #define vreinterpret_u32_s8
 
-uint32x2_t vreinterpret_u32_u64 (uint64x1_t t);
+_NEON2SSESTORAGE uint32x2_t vreinterpret_u32_u64 (uint64x1_t t);
 #define vreinterpret_u32_u64
 
-uint32x2_t vreinterpret_u32_s64 (int64x1_t t);
+_NEON2SSESTORAGE uint32x2_t vreinterpret_u32_s64 (int64x1_t t);
 #define vreinterpret_u32_s64
 
-uint32x2_t vreinterpret_u32_f32 (float32x2_t t);
+_NEON2SSESTORAGE uint32x2_t vreinterpret_u32_f32 (float32x2_t t);
 #define vreinterpret_u32_f32
 
-uint32x2_t vreinterpret_u32_p16 (poly16x4_t t);
+_NEON2SSESTORAGE uint32x2_t vreinterpret_u32_p16 (poly16x4_t t);
 #define vreinterpret_u32_p16
 
-uint32x2_t vreinterpret_u32_p8 (poly8x8_t t);
+_NEON2SSESTORAGE uint32x2_t vreinterpret_u32_p8 (poly8x8_t t);
 #define vreinterpret_u32_p8
 
-uint32x4_t vreinterpretq_u32_u16 (uint16x8_t t);
+_NEON2SSESTORAGE uint32x4_t vreinterpretq_u32_u16 (uint16x8_t t);
 #define vreinterpretq_u32_u16
 
-uint32x4_t vreinterpretq_u32_u8 (uint8x16_t t);
+_NEON2SSESTORAGE uint32x4_t vreinterpretq_u32_u8 (uint8x16_t t);
 #define vreinterpretq_u32_u8
 
-uint32x4_t vreinterpretq_u32_s32 (int32x4_t t);
+_NEON2SSESTORAGE uint32x4_t vreinterpretq_u32_s32 (int32x4_t t);
 #define vreinterpretq_u32_s32
 
-uint32x4_t vreinterpretq_u32_s16 (int16x8_t t);
+_NEON2SSESTORAGE uint32x4_t vreinterpretq_u32_s16 (int16x8_t t);
 #define vreinterpretq_u32_s16
 
-uint32x4_t vreinterpretq_u32_s8 (int8x16_t t);
+_NEON2SSESTORAGE uint32x4_t vreinterpretq_u32_s8 (int8x16_t t);
 #define vreinterpretq_u32_s8
 
-uint32x4_t vreinterpretq_u32_u64 (uint64x2_t t);
+_NEON2SSESTORAGE uint32x4_t vreinterpretq_u32_u64 (uint64x2_t t);
 #define vreinterpretq_u32_u64
 
-uint32x4_t vreinterpretq_u32_s64 (int64x2_t t);
+_NEON2SSESTORAGE uint32x4_t vreinterpretq_u32_s64 (int64x2_t t);
 #define vreinterpretq_u32_s64
 
-uint32x4_t vreinterpretq_u32_f32 (float32x4_t t);
+_NEON2SSESTORAGE uint32x4_t vreinterpretq_u32_f32 (float32x4_t t);
 #define  vreinterpretq_u32_f32(t) _M128i(t)
 
-uint32x4_t vreinterpretq_u32_p16 (poly16x8_t t);
+_NEON2SSESTORAGE uint32x4_t vreinterpretq_u32_p16 (poly16x8_t t);
 #define vreinterpretq_u32_p16
 
-uint32x4_t vreinterpretq_u32_p8 (poly8x16_t t);
+_NEON2SSESTORAGE uint32x4_t vreinterpretq_u32_p8 (poly8x16_t t);
 #define vreinterpretq_u32_p8
 
 //*************  Round ******************
-float32x4_t vrndnq_f32(float32x4_t a);
+_NEON2SSESTORAGE float32x4_t vrndnq_f32(float32x4_t a);
 #ifdef USE_SSE4
-#define vrndnq_f32(a) _mm_round_ps(a, _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC)
+#   define vrndnq_f32(a) _mm_round_ps(a, _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC)
 #else
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING( float32x4_t vrndnq_f32(float32x4_t a), _NEON2SSE_REASON_SLOW_SERIAL)
 {
@@ -16686,9 +16688,9 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING( float32x4_t vrndnq_f32(float32x4
 #endif
 
 
-float64x2_t vrndnq_f64(float64x2_t a);
+_NEON2SSESTORAGE float64x2_t vrndnq_f64(float64x2_t a);
 #ifdef USE_SSE4
-#define  vrndnq_f64(a)  _mm_round_pd(a, _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC)
+#   define  vrndnq_f64(a)  _mm_round_pd(a, _MM_FROUND_TO_NEAREST_INT | _MM_FROUND_NO_EXC)
 #else
 _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(float64x2_t vrndnq_f64(float64x2_t a), _NEON2SSE_REASON_SLOW_SERIAL)
 {
@@ -16703,10 +16705,10 @@ _NEON2SSE_INLINE _NEON2SSE_PERFORMANCE_WARNING(float64x2_t vrndnq_f64(float64x2_
 
 
 //************* Sqrt ******************
-float32x4_t vsqrtq_f32(float32x4_t a);
+_NEON2SSESTORAGE float32x4_t vsqrtq_f32(float32x4_t a);
 #define vsqrtq_f32 _mm_sqrt_ps
 
-float64x2_t vsqrtq_f64(float64x2_t a);
+_NEON2SSESTORAGE float64x2_t vsqrtq_f64(float64x2_t a);
 #define vsqrtq_f64 _mm_sqrt_pd